1111111111111111111111111111111111111111111111111111111111111111
. 3 1176 00160 3746
111111111111111111111111111111111111111111111 NF01145
NASA-CR-164604 19810022206
"",hi~ hook \'Ins jJI'ropnrc(i a..S an <leconnt of work sponsored by ~n 8gcllCY of the United : :i.".i:~s C:nvcnlment Neither the Unite(\ States Government nor allY agency thereof, nor Hny () f i hci( GUlpJOYCf;;J, ra~~k(~s vuy '\)J8xrnnty, express or hnpHed, or rrSSHrncs any legal liability or tCs!HH)~H)Hhy for iljc cH:cHn~ey, corflpletcnes~\~ or usefulness of any inforrnatioll, apparatu~;, ,IlJ'O(ll!Ct, or r)).oo,,'(; clisdo:;e(l, or represents that its usc would notinfringc privately oVillClj (i.vhi:~';""~;;.!;fnl'CHC0 1H!fCin 'in any ~pcc.ific conunercial product, proc0~;s, or service by tra(lc l.I;-l.~).H~, 'h>;:v:k:nwyk., 1!lPHHf~.ctl1.rer? or other\~J.he, (lOeB not necessarily constitut(~ or irnply its "",-d())~:nJ(]'~.nt, l'(:c"illr,1r;wb.1ion, or ffl1lorilli; by the: United S(;ltes Govemmcnt or any agency 1 hereof. THe vicYJ~ ;uA(l 01)iulons of (\uthon: expressed herein do not lleccr;sarily state or ,'(;{lcd. I ho:;c of t1~(; "OnHcd SI:Hi:f2:J (~overnn1ent or any (1ge~lCy thert;of."
1'\",jl;,Il\r; hm.11 ihr; N:ttiow.d 'l\:l;hnic~J Information Service, lU. S. Department of
:ndliJIf:"U;, ~)(lrinf\nckl, Vixginia /,:n6L
l'xke; Printed Copy AIO :MicrofichD AOI
PLANETARY BOUNDARY LAYER WIND MODEL EVALUATION AT A MID-ATLANTIC COASTAL SITE
H. W. Tie1eman
DOE/ET /23007-80/1 (DE81024093)
Distribution Category UC·60
Department of Engineering Science and Mechanics Virginia Polytechnic +nstitute and State University
B1acksbur~, Virglnla 24061
October 198.0
Prepared for the
U.S. Department: of Energy Office of Solar Power Applications
FEDERAL WIND ENE~GY PROGRAM
DOE Contract No. DE-AC06-79ET23007
This Page Intentionally Left Blank
ACKNOWLEDGEMENTS
F:lnancial assistance for preparation of this report from the
U. S. Department of Energy through Contract DE-AC06-79 ET 23007
is gratefully acknowledged. The development of the instrumentation,
data acquisition and data handling systems and the actual data
acquis:Ltion and data analysfs was supported by the National Aero
nautics and Space Administration, Grant NGL 47-004-067.
The continuous support of the grant officers Mr. J. F. Spurling,
W. H. lNest, R. E. Carr, F. Schmidlin and the Director of Wallops
Space :Flight Center Mr. R. L. Krieger is greatly appreciated. In
particular my thanks go to the technical assistance of Mr. C. A. Lewis,
J. Vs.n Overeem, P. H. Randall and R. L. Kelley without their help
the success of this research program would have been impossible. A
special thanks goes to Ms. E. Jane Harrison for her willing cooperation
and pe:rseverance in the typing of this report.
iii
This Page Intentionally Left Blank
EXECUTIVE SUMMARY
Detailed measurements of the mean flow and turbulence have been
made with the use of a micrometeorological facility consisting of an
instrumented 76-m ·tall tower located within a 100-m distance from
the Atlantic Ocean at Wallops Island, Virgi.nia. An interpretation of
the eJcperimental results demonstrates that under moderately strong
wind eonditions (hourly mean wind speed between 10 mls and 20 mls at
a height of 10 m), the popular neutral boundary-layer flow model faUs
to provide an adequate description of the actual flow.
For daytime westerly winds the convective boundary layer, which
has b~~en previously observed at sites on the continent, provides an
adequate model for the surface flow at the Wallops Island site.
Howev~~r variations from this model have been observed for certain
wind directions and under certain atmospheric conditions such as
low altitude cloud cover combined with precipitation. The observed
10w-flC'equency velocity fluctuations give ri.se to increased turbulent
intensities and larger turbulence integral scales. These low
frequlmcy fluctuations also occur in the surface layer where the
observed mean velocity profiles generally fit the logarithmic law quite
well.
For on-shore winds the surface flow is complicated as the re-
sult of the development of an internal boundary layer (IBL) as the air
cross:lng the beach generally experiences a change in surface roughness
and surface temperature. The internal boundary layer has a height
betwelm 15 m and 30 m at the tower location depending on wind direction
and change in surface conditions. For southerly winds the warmer alr
v
flows over the cooler water allowing the existence of a surface-based
inversion of variable depth. Under these conditions a low-altitude
maximum velocity (surface jet), occasionally below the highest
observation level of 76 m, has been observed. Under extreme stable
conditions at hourly mean velocities in excess of 10 m/s the tur
bulence has been observed to vanish completely. In addition, 10w
frequency internal gravity waves have been observed to co-exist
with the turbulence.
In addition to detailed flow information for all wind directions,
averages of the important flow parameters used for design such as
vertical distributjon of mean velocity, turbulence intensities and
turbulence integral scales have been presented for wind-direction
sectors with near-uniform upstream terrain. Power spectra of the
three velocity components for the prevailing northwesterly and southerly
winds are presented and discussed in detail.
The experimental results indicate clearly that the non-uniformity
of the upstream surface conditions, the non-neutral thermal stratifi
cation and the presence of appreciable low-frequency velocity fluctu
ations have a pronounced effect on the surface flow. Consequently it
is impossible to find a simple and single PBL model to describe the
flow at this site even under moderately strong wind conditions. More
over, there is no evidence that under still stronger wind conditions
(hourly mean wind speed at z=10 mover 20 m/s) the surface flow will
alter sufficiently as to conform to the neutral boundary-layer model
whose turbulence is of purely mechanical origin.
vi
TABLE OF CONTENTS
,Chapter
ACKNOWLEDGEMENTS •
EXECUTIVE SUMMARY . . . . . . • • • ., • • • • v
LIST OF' SYMBOLS . . . . ix
LIST OE' FIGURES . . . • • xi
LIST OF' TABLES . . . . . . . . . . . . · xix
1. INTRODUCTION · . 1
2. SITE DESCRIPTION • • 2
3.
4.
5.
6.
INSTRUMENTATION . . • . . 4
3.1 Cup-Vane Instruments • • 4
3.2: Hot-Film Anemometers • • • 5
DATA HANDLING AND DATA ANALYSIS • • • • • • 7
4.1 Cup-Vane Instruments • • • • • • • 7
4.2: Hot-Film Anemometers · 9
PLANETARY BOUNDARY LAYER (PBL) . 14
RESULTS AND DISCUSSION • . . . . • • • 22
6.1 Mean Wind and Mean Temperature Profiles •• • • • 23
6.2: Velocity Profile Parameters, z and a • o
6.3 Turbulence Intensities
• 28
• 33
Turbulence Integral Scales • • • 38
6.4.1
6.4.2
I 1 S 1 Lx LZ and LZ f C V D ntegra ca es, , rom up- ane ata u u v
Comparison of the Cup-Vane Scales with Predicted Values from References 25, 26 and 29 • • • • • •
6.4.3 The Direct and Spectral Methods for Obtaining
• 38
Turbulence Integral Scales • • • • • • • • • '+6
6.4~4 Comparison of Integral Scales Obtained Via the Direct and Spectrum Methods With Predicted Values of References 25, 26 and' 29 •••••••••••• 50
vii
Chapter
6.5 Power Spectra.
7. SUMMARY AND CONCLUSIONS
REFERENCES
FIGURES
TABLES •
viii
Page
55
62
69
74
183
Symbol
C
c P
f
g
K
k
k e
n
S
u
LIST OF SYMBOLS
Definition
Constant
Specific heat
Reduced frequency f=nz/U
Reduced frequency at the point of intersection of the extrapolated inert:ial subrange of the velocity spectra and the line nSi (n)/cri
2=1 for i=u,v and w~
Gravitational constant
Von Karman's constant
Wave number
Wave number range of energy-containing eddies
One-dimensional wave number
Monin-Obukhov length scale (4)
Streamwise turbulence integral scale for i=u,v,w (1).
Vertical turbulence integral scale for i=u and v (8)
Frequency
Surface heat flux
Autocorrelation function
Vertical correlation coefficient for i=u and v (9)
Gradient Richardson number (5)
Three-dimensional power spectral density function
One-dimensional power spectral density function for i=u,v and w
Time of integration and absolute temperature
Maximum time delay
Time
Mean Velocity
ix
Symbol
U*
U* o
u,v,w
v
x,y,z
z'
z o
8
e
A m
A o
p
T
T o
t
+
t
Definition
Local friction (shear) velocity, U* = [uw2 + vw2]1/4
Friction (shear) velocity obtained from the logarithmic velocity profile (2)
Reference velocity
Turbulence velocity components in x,y and z directions respectively
Magnitude of mean velocity
Mean wind coordinates
Vertical separation distance
Roughness length (2)
Reference height
Power law exponent of mean velocity profile (6)
Adiabatic lapse rate
Absolute temperature
Temperature fluctuations
Wavelength
Wavelength associated with spectral peak
Wavelength associated with f o
Air density
Standard deviation of the fluctuating velocity compoent i=u,v and w
Time delay
Surface stress
Wind direction
Universal function of z/L (3)
Upward integration for one-sided integral scales (8, 9)
Downward integration for one-sided integral scales (8,9)
Two-sided integration for two-sided integral scales (10, 11)
x
LIST OF FIGURES
FIGURE
1. Map of Chesapeake Bay and Delmarva Peninsula • • • • • • • 75
2. Wallops Island, Immediate Surroundings of the Meteorological Tc)wer . . . • . . . • . • . . . . . . • . . • • . 76
3. W:lnd Direction Sectors wj.th Upstream Terrain Features 77
4. The 76.2m (250 ft) Micrometeorological Tower at Wallops Island. 78
5.
6.
7a.
7b.
7c.
7d.
8.
9.
10.
11.
12.
13.
14.
15a.
Afternoon Temperature Profiles for Southerly Winds •
Temperature Profiles for Westerly Winds · · · · · · · · · · · Variation of Mean-Velocity Ratio, V(15. 2) /v(76. 2) with Wind Direction · · · · · · · · · · · · · · · · · · · · · · · · · · Variation of Mean-Velocity Ratio, v(30. 5) /V(76. 2) with Wind D:lrection · · · · · · · · · · · · · · · · · · · · · · · · Variation of Mean-Velocity Ratio, V(45.7/V(76.2) with Wind D:lrection · · · · · · · · · · · · · · · · · · · · · · · · Variation of Mean-Velocity Ratio, V(61. 0) /V(76. 2) with Wind D:lrection · · · · · · · ' . · · · · · · · · · · · · · Typical Strong-Wind Profiles for Southerly Wind Directions · · Typical Strong-Wind Profiles for Westerly Wind Directions Gradient Richardson Numer (5) Evaluated at z=15.2m •••••
Variation of the Roughness Length, z , Versus Wind Direction Note the Larger Roughness Lengths inOthe Sector 350 0 <<1><+30 0
Variation of the Roughness Length, z , with the Gradient o R:lchardson Number (5) Evaluated at z=15. 2m • • • • • • •
Variation of Power-Law Exponent with Roughness Length, Both Based on Velocity Measurements Below 45.7m. Panofsky's RI~lation (7) Based on zl=:15.2m and z2=45.7m •••••••••
Variation of Power-Law Exponent with Roughness Length, Both Based on Velocity Measurements at all Levels. Panofsky's RI~lation (7) Based on zl=15.2m and z2=76.2m •••••••••
Variation of Power-Law Exponent with Roughness Length, Both Based on Velocity Measurements Above 45.7m. Panofsky's RI~lation (7) Based on zl=:45.7m and z2=76.2m
Panofsky [31] --_. Counihan [26] ___ ESDU [27] ---Variation of Turbulence Intensity, (Ju/U, with Direction, z==15.2m • • • • . • .. • • • • • 0 · ., . . ~ . . .
xi
79
80
81
82
83
84
85
86
87
88
89
90
91
92
FIGURE PAGE
1Sb. Variation of Turbulence Intensity, a Iu, with Direction, z=30.Sm u 93 · · · · · · · · · · · · · · · · · · · · · · · · · · ·
15c. Variation of Turbulence Intensity, au/u, with Direction, z=47.5m · · · · · · · · '. · · · , . · · · · · · · · · · · · · · 94
15d. Variation of Turbulence Intensity, a IU, with Direction, z=61.Om u
95 · · · · · · · · · · · · · · · · · · · · · · · · · · · 15e. Variation of Turbulence Intensity, a Iu, with Direction,
z=76.2m u 96 · · · · · · · · · · · · · · · · · · · · · · · 16a. Variation Turbulence Intensity,av/U, with Direction,
z=15.2m · · · · · · · · · · · · · · · · · · · · · · · 97
16b. Variation Turub1ence Intensity, a Iu, with Dl.rection, z=30.5m v 98 · · · · · · · · · · · · · · · · · · · · · · · ·
16c. Variation Turbulence Intensity, av/u, with Direction, z=47.5m · · · · · · · · · · · · · · · · · · · · · · · · · · · 99
16d. Variation Turbulence Intensity, av/u, with Direction, z=61.Om · · · · · · · · · · · · · · · · · · · · · · · · · 100
16e. Variation Turbulence Intensity, a lu, with Direction, z=76.2m v 101 · · · · · · · · · · · · · · · · · · · · · ·
17. Variation of Average Turbulence Intensities a lu and a lu u v with Height for Two Wind-Direction Sectors • • • • • • • • • 102
18. Typical Stirong-Wind Aerovane Record of Direction and Speed for Northwest Winds, CP~320°, U10~13 m/s ••••••••••• 103
19. Probability Density Function of Wind-Direction Fluctuations, z=76.2m, acp=8°, Cp=322° ••••••••••••••••• 104
20. The Variation of the Vertical Distribution of Turbulence Intensity au/u, with Time of Day for Westerly Winds • • • • • lOS
21. Vertical Distribution of Turbulence Intensity, a" lu, for Westerly Winds • • • • • • • • • • • • • ~ • • • • 106
22. Vertical Distribution of the Turbulence Intensity, a IU, for Southerly Winds u 107 · · · · · · · · · · · · · · · · · · · · · · ·
23. Vertical Distribution of the Turbulence Intensity, av/u, for Southerly Winds · · · · · · · · · · · · · · · · · · · · · · · 108
24. Vertical Distribution of'the Turbulence Intensity, aw/U, for Southerly Winds · · · · · · · · · · · · · · · · · · · · · · 109
25. Variation of Ratios of Average Turbulence Intensity with Wind Direction · ,.. . . . . . . . . . . . . . . . . 110
xii
FIGURE PAGE
26a. Variation of Turbulence Integral Scale, x Lu with Wind
Direction, z=15.2m · · · · · · · · · · · · · · · · · · • 111
26b. Variation of Turbulence Integral Scale, LX with Wind Direction, z=30.5m u' 112 · · · · · · · · · · · · · · ·
26c. Variation of Turbulence Integral Scale, LX with Wind Direction, z=45.7m.
u' 113 · · · · · · · · · · · · · · 26d. Variation of Turbulence Integral Scale, LX with Wind
Direction, z=61. Om u' 114 · · · · · · · · · · · · · · ·
26e. Variation of Turbulence Integral Scale, LX with Wind Direction, z=76.2m •
u' 115 · · · · · · · · · · · · 27a. Variation of Turbulence Integral Scale, LZt with Wind
Direction, z=15.7m u ' 116 · · · · · · · · · · · · · · · · · · 27b. Variation of Turbulence Integral Scale, LZt with Wind
Direction, z=30.5m u ' 117 · · · · · · · · · · · · · · · · · · 27c. Variation of Turbulence Integral Scale, LZt with Wind
Direction, z=45.7m u ' 118 · · · · · · · · · · · · · · · · · · 27d. Variation of Turbuelnce Integral Scale, L~t, with Wind
Direction, z=4S.7m · · · · · · · · · · · · · · · · · · 119
27e. Variation of Turbulence Integral Scale, LZ+ with Wind Direction, z=45.7m. u ' 120 · · · · · · · · · · · · · · · · ·
27f. Variation of Turbulence Integral Scale, LZ+ with Wind Direction, z=61. Om u ' 121 · · · · · · · · · · · · · · · · · ·
27g. Variation of Turbulence Integral Scale, LZ+ with Wind direction, z=76.2m u ' 122 · · · · · · · · · · · · · · · · · ·
28a. Variation of Turbulence Integral Scale, LZt with Wind Direction, z=15.2m v ' 123 · · · · · · · · · · · · · · · · · ·
28b. Variation of Turbulence Integral Scale, LZt with Wind Direction, z=30.5m v ' 124 · · · · · · · · · · · · · · · · · · · · ·
28c. Variation of Turbulence Integral Scale, LZt with Wind Direction, z=45.7m v ' 125 · · · · · · · · · · · · · · · · · · · · ·
28d. Variation of Turbulence Integral Scale, LZt with Wind Direction, z=45.7m v ' 126 · · · · · · · · · · · · · · · · · · · · ·
28e". Variation of Turbulence Integral Scale, LZ+ with Wind Direction, z=45.7m v ' 127 · · · · · · · · · · · · · · · · · · · · ·
xiii
FIGURE PAGE
28f. of Turbulence Z
with Wind Variation Integral Scale L +, Direction Z = 61. Om • v 128 . . · . . . . . . . . . . . . . · ·
28g. Variation of Turbulence Integral Scale, LZ+ with Wind Direction z=76.2m v ' 129 . · . . . · ·
29. Variation of the Integral-Scale Ratio LX/Lz for the Three Lowest Observation Levels with Wind Direction. u A¥rows Indicate the Direction of Integration • • 130
XI Z 30. Variation of the Integral-Scale Ratio L LU for the Three Highest Observation Levels with Wind Di¥ection. Arrows
31.
32.
33.
34.
35.
36.
37.
38.
39.
Indicate the Directicn of Integration • • • • • • • • • • • • 131
X Variation of Aever~ge Integr~l Scales Lu(Cup-Vane) with Height for Two Westerly W1nd-Direct10n Sectors • • • • • • • • • • • 132
. Z Variation of Average Integral Scales, L , (Cup-Vane) with Height for Two Westerly Wind-Direction Sectors~ Arrows on Symbols Indicate the Direction of Integration ••••••••••• 133
Variation of Average Integral Scales, LZ, (~up~Vane) with Height for Two Westerly Wind Direction Sectors~ Arrows on Symbols Indicate the Direction of Integration ••••••••••• 134
Variation of Average Integral Sca1e,Lx , (Cup-Vane) with Height for Wind-Direction Sector 180°-195° (Sguth Winds) z<2Om+zo=0.01m and z>20m+z =O.OOlm ••••••••••••••••••••• 135 o
Variation of Average Integral Scales, L z, (Cup-Vane) with Height for Wind-Direction Sector 180°=195° (Sguth Winds). z<2Om+z =O.Olm and z>2Om+z =O.OOlm. Arrows on Symbols Indicate
o 0 the Direction of Integration. • • • • •• •• 136
Variation of Average Integral Scales, L z, (Cup-Vane) with Height for Wind-Direction Sector 180°-195° ($o~th Winds) z<2Om+z=Q.01m and z>2Om+20m=0.001m. Arrows on Symbols Indicate the Direction of Integration • • • • • • • • • • • • • • • • • • • • • 137
Profiles of Average Integral Scales, LX, for Westerly Winds u
for Different Times of the Day • • • • • • • • • 138
Variation of the Integral Roughness Length, z , for Direction and Ri 15>£0.1 •
X Scale, Lu ' at z=15.2m (50 ft.) with all Data Records with a Westerly Wind · . . . . . . . . . . • . . . . . . 139
x Variation of the Integral Scale, L , at z=45.7m (150 ft.) with . u Roughness Length, z , for all Data Records with a Westerly Wind Direction and ~i15>-0.1 •••••••••••••••• 140
xiv
FIGURE
40. Variation of the Integral Scale, LZt, at z~45.7m (150 'with Roughness Length, Z , for allUData Records with Westerly Wind Direction gnd RilS>-O.l ••••••••
PAGE
ft.)
141
41. 'Variation of Integral Scale, L2t, at z=4S.7m (150 ft.) with Roughness Length, z , for all ~ata Records with Westerly Wind Direction and fh15>-0.1 ••••••••.••••••• 142
42. Logarithmic u-Spectra for Run 19, a .and LX Obtained in the Middle Frequency Range, LX ObtaineduVia th~ Direct Method. • 143
u
43. Logarithmic u-Spectra for Run 19, a and LX Obtained in the lffigh Frequency Range, LX Obtained V¥a the uDirect Method 144
u
44. Logarithmic w-Spectra for Run 19, a and LX Obtained in the Middle-Frequency Range, LX ObtainedwVia th~ Direct Method. • 145
w
45.
46.
][,ogarithmic w-Spectra for Run 19, a Frequency Range, LX Obtained Via th~
and LX Obtained in the High-
w Direcf Method 146
Variation of Average Integral Scales, LX, Obtained with the Hot-Film System (Von Karman Method, HigH and Middle Frequency Range) and with the Cup--Vane System (Data of the Two Systems Taken Simultaneously), with Height for Southerly Winds 147
47. Variation of Average Integral Scales, LX, Obtained with the Hot-Film System (Von Karman Method, HigK and Middle Frequency Range) with Height for Southerly Winds • • • • • • • . 148
48. Variation of Average Integral Scales, ~, Obtained with the Hot-Film System (Von Karman Method, High and Middle Frequency Range), with Height for Southerly Winds ••••••.• 149
50.
51.
Variation of Average Integral Scales, LX, Obtained with the Hot-Film System (Von Karman and Direct Methods, Middle-Frequency Range) and Cup-Vane System with Height for Northwesterly winds . . . . . . • . . . . . . . . . . . . . 150
Variation of Average Integral Scales, LX, Obtained with the Hot-Film System (Von Karman and Direct :Methods, Middle Frequency Range) with Height for Northwesterly Winds • • • • . . 151
Variation of Average Integral Scales, LX, Obtained with Hot-Film System (Von Karman and Direct WMethods, Middle l~requency Range) with Height for Northwesterly Winds
xv
the
152
FIGURE PAGE
52. Comparison of Integral Scales, LX, of Record #16. (Evening Run) (Von Karman and Direct Methgds, Middle Frequency Range) Against Average of Daytime Hot-Film Runs (Von Karman, Middle Frequency Range) for Northwesterly Winds • • •• •••• 153
53. Comparison of Integral Scales, LX, Obtained from the Hot-Film System and Analyzed in the Middlg-and High-Frequency Range Using the Von Karman Method for Northwesterly Winds • • .' l54
54. Time Records of Velocity Components. Elapsed Time 55
55.
Minutes
Comparison of Integral Scales, LX, of Record #16 (Evening Run) (Von Karman and Direct Meth'Zds, Middle Frequency Range) Against Average of Day Time Hot-Film Runs (Von Karman, Middle
+55
Frequency Range) for Northwesterly \'linds • • .• •••• 156
56.
57.
58.
.Comparison of Integral Scales, LX,. Obtained From the Hot-Film System and Analyzed in the Middl~-and High-Frequency Range Using the Von Karman Method, for Northwesterly Winds • • •• 157
X Comparison of Integral Scales, Lw' of Rec;:ord 1116 (Evening Run) (Von Karman and Direct Methods, Middle Frequency Range) Against Average of Day Time Hot-Film Runs (Von Karman, Middle Frequency Range) . • • • • • • • • • • •• 158
Comparison of Integral Scales, LX, Obtained From the Hot-Film System and Analyzed in the ~iddle- and High-Frequency Range Using the Von Karman Method for Northwesterly Winds. ~ 159
59a. Logarithmic u-Spectra Versus Modified Reduced Frequency, flf , o Analyzed in the High-Frequency Range, South Winds, Unstable Thermal Stratification. -l.O<z/L<O • • • • • • • • • • • 160
59b. Logarithmic u-Spectra Versus Modified Reduced Frequency, flf , Analyzed in the High-Frequency Range, South Winds, Stable 0
Thermal Stratification. O<z/L<+l.O •••••••••• 161
59c. Logarithmic u-Spectra Versus Modified Reduced Frequency, flf , Analyzed in the High-Frequency Range, South Winds, E~tremelyO Stable Thermal Stratification. z/L>+l.O • • • • • • • • 162
60a. Logarithmic v-Spectra Versus Modified Reduced Frequency, flf , Analyzed in the High-Frequency Range, South Winds, Unstable 0
Thermal Stratification. -1. O<z/L<O • • • • • • • • • • • 163
60b. Logarithmic v-Spectra Versus Modified Reduced Frequency, flf , Analyzed in the High-Frequency Range, South Winds, Stable 0
Thermal Stratification, O<z/L<+1. 0 • • • • • • • • • • • 164
60c. Logarithmic v-Spectra Versus Modified Reduced Frequency, flf , Analyzed in the High-Frequency Range, South Winds, ExtremelyO Stable Thermal Stratification, z/L>+l.O • • • • • • • • •• 165
xvi
FIGURE PAGE
61a.
61b.
61c.
Logarithmic w-Spectra Versus Modified Reduced FrequencYt flf t Analyzed in the High-Frequency Range t South Winds t Unstable 0
Thermal Stratification, .-1. O<z/L<O • • • • • • • • • • • • • 166
Logarithmic w-Spectra Versus Modified Reduced Frequency, flf , Analyzed in the High-Frequency Range, South Winds, Stable 0
Thermal Stratification, O<z/L<+1.0 • • • • • • • • • • • • • 167
Logarithmic w-Spectra Versus Modified l~educed Frequency, flf , o Analyzed in the High-Fr.equency Range, South Winds, Extremely
Stable Thermal Stratification z/L>+1. 0 • • • • • •• •• 168
62. The Vertical Distribtu:lon of the Reduced Frequency (f ) i With i=u,v,w for Southerly Winds. Data Analyzed in 0
the High-Frequency Range 0.0244<n<100 Hz ••••••••••• 169
63. Logarithmic u-Spectrum Versus Reduced Frequency, Analyzed in the Middle-Frequency Range, South Wind, Record #7, z=9 . 14m ...................,...... 170
64. Logarithmic u-Spectrum Versus Reduced Frequency, Analyzed in the Middle-Frequency Range. Southwind, Record #7, z=45.7m ..... tI • • • • • • • • • • • • • • • • • • • • 171
65. Logarithmic u-Spectra Versus Modified Reduced Frequency, flf , o Analyzed in the High and Middle Frequency Range, Northwest Wind, <1>=301° ••• 0 •••••••••••••••••••• 172
66. Logarithmic v-Spectra Versus Modified Reduced Frequency, flf , o Analyzed in the High and Middle Frequency Range, Northwest Wind, ¢=301o
•• • It • • • • • • • • • • • • • • • • • • • • 1.73
67. Logarithmic w-Spectra Versus Modified Reduced Frequency, flf , o Analyzed in the High and Middle Frequency Range, Northwest
Wind, <1>=301° ••• • • • • • • • • • • • • • • • • • • • • • 174
68. Logarithmic u-Spectra Versus Modified Reduced Frequency, flf , o Analyzed in the High and Middle Frequency Range, Evening
Run, Northwest Wind, <1>=292° •• 0 • • • • • • • • • • • • • • 175
69. Logarithmic v-Spectra Versus Modified Reduced Frequency, flf , Analyzed in the High and Middle Frequency Range, Evening 0
Run, Northwest Wind <1>=292° •••••••••••• • • • • • 176
70. Logarithmicw-Spectra Versus Modified Reduced Frequency, flf , Analyzed in the High and Middle Frequency Range, Evening 0
Run, Northwe'st Wind, <1>=292° • • • • • • • • • • • • • • • • • 177
71. The Vertical Distribution of the Reduced Frequency (f )i with i=u,v,w for Westerly Winds. Data Analyzed in thg Middle-Frequency Range 0.0015<n<6.25 Hz. Dark Symbols Denote the Data for the Evening Run 16 • • • • • • • • • • • • • • • 178
xvii
FIGURE PAGE
72. Average Velocity Ratios, V/V250 for Each Wind-Direction Sector with Near-Uniform Upstream Roughness (See Figure 3) . . . . . . . . . . . . . . . . . . . . .,. . 179
73.
74.
75.
Average Turbulence Intensity,ouIU, for Each Wind-Direction Sector with Uniform Upstream Roughness (See Figure 3) • • •
Average Turbulence Intensity, ovlU, for Each Wind-Direction Sector with Near~Uniform Upstre~ Roughness (See Figure 3 .
x Average Turbulence Integral Scales, L , for Each Wind
u Direction Sector with Near-Uniform Upstream Roughness (See Figure 3) ................... .
xviii
180
181
182
LIST OF TABLES
TABLE PAGE
1. Mean Profile and Turbulence Parameters Over Near-Uniform Terrain . . . . . . . . . . . . . . . ~ . . . . . . . . . . . . 184
xix
This Page Intentionally Left Blank
1. INTRODUCTION
The purpose of this report is to provide information on the local
wind climate at a mid-At1ant:Lc coastal site. The acquired information
can ble used for the design of wind-turbine generators at similar sites.
Since wind is a very important design parameter for these generators,
information is provided in this report on wind speed, wind direction,
wind shear and wind turbulence.
The data presented in this report were collected from an instru
mented meteorological tower, 76.2 m (250 flaet) tall and located at
Wallops Island. This island is one of the barrier islands at the
Atlantic coast along the Eastern Shore of Virginia and is used by
the National Aeronautics and Space Administration as a sounding rocket
launch facility. The results acquired frOln this facility should be
typical for any Atlantic coastal site, although local effects such as
upstream buildings and obstacles and changes in surface roughness and
surface temperature modify the flow near the surface. At Wallops
Island, the surrounding terrain beyond a distance of 100-300 m from
the tower can be considered as homogeneous and uniform, so that the
flow above a height of approximately 10-30 m should not be affected
by local terrain nonuniformities.
The wind and temperature data from this site were acquired under
moderately strong wind conditions with an hourly mean velocity of at
least 10 mls at the 76.2 m(250 ft) level. For wind directions be-
tween northeast and southwest this requirement had to be reduced to
approximately 8 mis, since strong winds from this sector occur very seldom.
2
Mean wind and turbulence measurements were made with two types of
instrumentation consisting of cup-vanes and resistance temperature
probes primarily used for mean profile measurements of velocity and
temperature respectively. In addition, the cup-vane instruments were
used for turbulence intensities of the two horizontal velocity compo
nents and horizontal and vertical turbulence integral scales. The hot
film and thermocouple system was used for measurement of turbulence
intensities, turbulence fluxes and velocity spectra in all three
directions. The cup-vane system was used to acquire wind data from
all directions, while the hot-film system was only used for turbulence
measurements from the two prevailing wind directions, south and north
west.
The results of this experimental research are presented in a form
suitable for design purposes. Where ever possible the results are also
compared with previously published results and with existing empirical
models for near-neutrally stratified low-level winds.
2. SITEDESCRIPTION
Wallops Island consists of a narrow strip of dunes, appr9ximately
3 meters above sea level, and is situated in a northeast-southwest
direction. The island is separated from the "Delmarva" peninsula by
a tidal marsh on the west side, and with the Atlantic Ocean on the east.
Winds with directions varying between west and north are usually
encountered following the passage of a cold front. Winds from this
sector will have crossed in succession (Fig. 1) the mainland, the
Chesapeake Bay (20-50 km), the "Delmarva" peninsula (20-50 km) and the
tidal marsh (3-5 km). Depending on the wind direction, for the last
3
200-300 m the air travels over land before it arrives at the tower
location (Fig. 2). The tidal marsh between the island and the
peninsula consists of shallow areas of water interchanged with swamp
vegetation, mostly grass of a maximum height of 1 m. Some taller
vegetation consisting of bushes and brush of a maximum height of 5 m
exists in several upstream d:lrections. For wind directions between
255° and 270°, a 6.5 m high rocket fuel storage bunker is approximately
90 m upstream (Fig. 2). An elevated roadway (levee) 2 m above the
surrounding terrain passes the tower on the west side within 200 m.
Winds with directions varying between north-east and south approach
the island from the Atlantic Ocean.
Sectors with approximately the same immediate upstream roughness
have been established as shown in Fig. 3. Between 0° (north) and
30° the upstream terrain features two bunkers within a distance of
100 m from the tower. In addition, a few small buildings and inter
mittent patches of brush are upstream as far as 750 m. Between 30°
and 45° (wind direction parallel to the island) many buildings are
upstrea.m and winds in this sector should encounter the roughest terrain
at this site over a distance of approximately 4 km. For wind directions
between 45° and 210° the winds approach the island over water, and cross
the beach at varying distances from the tower depending on the direction.
For wind directions between 1400 and 170° a one-story rocket assembly
building is about 100 m upstream of the tower. The prevailing southerly
winds vary in direction between 1700 and 210°, however in the sector
between 195 0 and 210° the 45-m tall Aerobee tower and associated
buildings are about 300 m upstream. For directions between 210 0 and
230°, the wind direction is approximately parallel to the island
4
with part of the Aerobee tower complex approximately 300 m upstream.
In this same sector a few other buildings, levees and sand dunes are
upstream at greater distances. Between 230° and 330° the upstream
terrain is very uniform with no big obstacles other than the afore
mentioned bunker and roadway. For the sector between 330° and 360°,
several patches of brush, 2 levees and one radar building are upstream
of the tower within a distance of 500 m with marsh at further upstream
distances.
3. INSTRUMENTATION
3.1 Cup-Vane Instruments
The 76 m (250 ft) micrometeorQlogical tower is a self-standing non
guyed tower with working platforms at 15.2 m (50 ft) intervals (Fig. 4).
The cup-vane velocity-direction instruments and aspirated temperature
probes both primarily used for profile measurements are mounted at 5
levels near each platform. Two sets of cup-vane instruments are mounted
at each level on 2m booms on opposite sites of the tower (Fig. 4).
An automatic electronic switching circuit ensures that data are taken
only with the instruments on the upwind side of the tower. The
electronics associated with this instrumentation system, together with
a digital readout panel of all instruments from one side of the tower,
are located in a small instrumentation building at the base of the
tower (Fig. 4). From this location the digitized data are transmitted
to the NASA control center at the main base on the peninsula about
13 km to the northwest. Here the data from each level sampled at a
rate of 1 sample each 2 seconds are recorded on digital tape. At
this sample rate, data can be acquired without interruption for about
5
8 hours. This instrumentation system is used by NASA in conjunction
with its rocket launching operations. Regularly scheduled mainten
ance and calibration of this system are performed by personnel under
NASA's supervision.
3.2 Hot-film Anemometers
Six three-dimensional split-film anemometers (TSI-10S0D) are
used for turbulence measurements, which include turbulence intensit:les,
turbulence fluxes, spectra and cross spectra of all three turbulence
components and temperature. These anemometer systems were chosen for this
research program since they have the advantage of small physical size,
fast response and high sensitivity over a wide range of velocities. The
instruments are mounted on 1.S m booms at the same levels as the cup
vane instruments and also at the 9 m (30 ft) level. Each hot-film probe
is mounted on a rotor, which is capable of rotating the probe about a
verti.ca1 axis so as to align the probe axis approximately into the mean
wind direction. The probe·-rotor combination is mounted on a 1.8 m-boom,
which in turn is mounted 011 the railing at each platform. The probes
were mounted on the south side of the tower for measurement of the
preva.i1ing south winds during the summer and on the north side of the
tower for measurement of the prevailing northwest winds during the
winter and spring. The electronics as well as the data-acquisition
and data-handling system for this instrumentation system are located
in an instrumentation trailer parked at the base of the tower (Fig. 4).
Each hot-film probe consists of three split-film sensors used for
measurement of wind speed and direction and a copper-constantan
thermocouple used for temperature measurements. Each sensor consists
of a 0.15 mm diameter quartz rod coated with a platinum film of about
6
1000 angstrom in thickness. The platinum film on each rod consists
of two segments, separated from each other by two longitudinal splits
1800 apart. The active elements on each rod are electrically heated
to the same constant temperature by separate anemometer circuits.
The total sensor length is about 5 rom, and the three sensor5 are mounted
mutually perpendicular to form a Cartesian coordinate system. When the
instruments are not used for data acquisition, the three sensors and
thermocouple are protected by an aluminum shield which can be moved
pneumatically to cover the sensors. As an added precaution, dry fil
te~od air is allowed to blow across the sensors when the shield covers
the sensors. This is done to protect the sensors from contamination in
the salt-air environment and moisture while not in operation. For a
more detailed review of the hot-film anemometer system th~ reader is
advised to consult Reference 1.
Calibration of the hot-film anemometers is carried out in a low
speed wind tunnel located at the main base. In order to obtain data of
a desired accuracy from the hot-film instrumentation system, a new
calibration and operating procedure was developed. Instead of using
the calibration constants supplied by the manufacturer, all constants
were obtained from calibration procedures carried out in the low-speed
wind tunnel and thermal chamber. This procedure proved to be both time
consuming and complicated but necessary. Calibration of each instrument
in the wind tunnel was carried out for 11 wind approach angles between
plus and minus 50 0 and for 13 velocities in a range varying between
0.3 and 15 mise The best accuracy of the data was obtained for wind
7
directi.ons parallel to the axes of the instrument, and consequently
the tower mounted instruments were rotated in the direction of the
mean wi.nd before the data acquisition was started. For details of
the calibration procedure and the relations for the conversion from
output voltage to velocity components it is suggested that the reader
consult References 1 and 2.
4. DATA HANDLING AND DATA ANALYSIS
4.1 Cu.p-vane Instruments
The output signals from the cup-vane instruments and temperature
sensing probes are sampled and digitized at a rate of 1 sample per
second. This information is transmitted to the control center of
the mai.n base, where every other sample is recorded on digital tape.
The data from these tapes, each capable of storing up to 8 hours of
data, IlLre then analyzed on the HW-625 computer at NASA, Wallops Flight
8 Center. The data are analyz~d in blocks of 2 =256 samples, representing
a data record of 512 seconds. For each l3ample the east-west and north-
south velocity components are calculated and averaged over 256 samples
from which the mean velocity and the mean direction for each block are
obtained. Also a block mean for the temperature is calculated.
Reasonably stationary sample records of 5 to 10 blocks in length are
selected for further analysis. This selection is based on the inspection
of the printout of the block means of velocity, direction and temperature
for all five levels.
Next the east-west and north-south velocity components and tempera-
ture are averaged for the selected sample, from which the sample mean
velocity and sample mean direction are calculated. This direction
8
defines the mean-wind coordinate system with the x-axis parallel to the
direction of the sample mean wind, the y-axis in the horizontal plane
perpendicular to the x-axis and the z-axis vertically upward. For all
the data points in each block the velocity components in the mean-wind
coordinate system are calculated and averaged to obtain the block means.
After the block means were removed from each set of components, variances
and covariances are calculated for each block. Sample variances and
covariances are obtained by averaging of the block variances and covariances
over the total number of blocks. The. covariances calculated in this manner
include all the combinations of like velocity components at the different
levels, allowing for the calculation of the vertical turbulence integral
scales of both the u and v components. In addition, the autocorrelation
function, R (T) of the streamwise velocity is calculated from which the u
turbulence integral scale, L~, is obtained as follows:
R (T)dT, U
(1)
where Tl is the time delay for which the first zero-crossing of the
calculated autocorrelation function occurs. The turbulence data
acquired with the cup-vane system are analyzed in a limited frequency
range of 0.00195-0.25 Hz.
A total of 195 digital data tapes were generated during the period
of July 1974 and December 1978. Approximately 300 data samples were
9
analyzed each varying between 43 and 85 minutes. Initially data were
acquired with the cup-vane system only, it was not until February 1976
that the temperature system came on line. However this system is not
too rel:iable and often temperature at one or two levels is missing
as the result of the equipment being down or out of calibration. Before
a lightning-arrester system was installed on the tower, excessive
amount of damage was inflicted on all systems during thunderstorms as
a result of line power surges and voltage induction in the cables that
(!onnect the instruments on the tower to the electronics at the base of
the tower. During the summer of 1976 a thermograph for recording the
,air temperature at ground level was added to the system.
Oceasionally when the equipment on the 76m (250 ft) tower was down,
data-acquisition was switched. to the 91 m (300 ft) tower located at the
north end of the island. This tower is instrumented with cup-vane
systems at six levels but has no temperature instruments. Its location
from thE~ beach is 280 m as compared to the 76 m (250 ft) tower which is
approximately 150 m from the beach. No major buildings or other obstacles
I;!xist bE~tween the 91 m (300 ft) tower and the beach. However, for ocean
winds the overland distance is longer and more modification of the undis
turbed ocean winds can be expected at the 91 m (300 ft) tower.
1+.2 Hot-film Anemometers
The data-acquisition and data~hand1ing system is designed to
handle output from six split-film anemometer systems, sampled at a rate
of 200 samples per second for a period of approximately one hour.
This system consists of two main parts: (a) the multiplexing and
~ma10g recording system and (b) the demu1tip1exing, digitizing and
10
digital recording system. The seven output voltages from each anemo~
meter are frequency modulated by voltage-controlled oscillators each
with a different center frequency. There is one set of voltage-con
trolled oscillators for each probe. The seven frequency-modulated
signals together with a 100 kHz reference signal are fed into a summing
amplifier to produce one single multiplexed signal. The multiplexed
signals from each instrument are recorded on separate channels of an
analog tape recorder together with time-of-day, which serves as a
reference for the recorded data.
At a later time, each of the multiplexed signals is demultiplexed
into its seven analog components after passage through seven discri
minators. In order to avoid aliasing of the velocity spectra the
six output voltages corresponding to the six split films are passed
through a 100 Hz low-pass filter. Next the analog voltages are sampled
at a rate of 200 Hz, digitized and recorded on digital tape.
A mini-computer (DEC Model PDP 11/20) controls the multiplexing analog
to-digital conversion and the digital recording. Access to the mini
computer is obtained with a teletypewriter. The data conversion starts
at a time-of-day prescribed by the operator, and the analog-to-digital
converter performs successive scans and conversions of seven analog vol
tages into 16 bit words at a rate of one scan each 5 milliseconds.
These words are stored in one of the buffers of the mini-computer which
in turn transfers the data to a 9-track digital magnetic tape. Each
buffer has a capacity of 209 scans representing 1.05 seconds of data.
A total of 3300 records make up a single sample record over a time
period of slightly less than one hour. 'The tapes with the digitized
11
data are taken to VPI and SU where the data are analyzed on an IBM-370
computer. Four separate computer programs have been developed to
calculate the following major statistical parameters: mean values,
variances, covariance spectra and cross spectra.
The first step in the data-analysis procedure is to convert the
seven output voltages from each film to three velocity components in.
the sensor-oriented coordinate sYEitem and temperature, using the con
stants obtained from the calibration data. The converted data are
transferred on another magnetic tape to await the next step of the data
reduction.
In the second program velocity and temperature data are analyzed
in blocks of N=2l3=8l92 data points, representing nearly 41 seconds
of data. For each of these blocks of data mean velocity components,
mean velocity and direction, mean temperature and the four standard
deviations are calculated. A total of 80 data blocks (almost 55
minutes) are analyzed in this manner. A stationarity trend test is
performed on each of the calculated parameters to check for un
acceptable nonstationarities. Also inspection af the printout of the
block parameters helps in the decision whether or not to continue with
the statistical anlaysis. At this point blocks with unrealistic data
can bE~ recognized and omitted from the data sample in future analysis.
The sample mean velocity components are obtained by averaging the block
means, allowing the calculation of the horizontal angle between the
sample~ mean-wind direction and the probe axis. In the following step
this angle is needed to tranfer the original velocity components in
the sensor-oriented coordinate system into u, v and w velocity components
of the mean-wind coordinate system as defined previously.
12
Block means are calculated for the temperature and for the velocity
components in the mean-wind coordinates system and removed from the data
in each block. The resulting fluctuating components are recorded on
magnetic tape for further analys~s. Also variances and all covariances
for each block are calculated and averaged over 80 blocks to obtain the
sample variances and covariances. The statistical parameters, including
the spectra to be calculated in the next step, contain only contributions
from the fluctuations in the frequency range of 0.0244-100 Hz. In order
to include contributions from frequencies below 0.0244 Hz, sixteen con
secutive data points are averaged into one data point to form a new
data record, which is also recorded on magnetic tape. This averaging
is performed after the data are transformed into the mean-wind coordi
nates and before the block means are removed. In this way only 5 blocks,
each 10.92 minutes long, are analyzed allowing for data analysis in the
frequency range of 0.00153-6.25 Hz. For these new data records, block
variances and covariances and sample variances and covariances are also
calculated. For the lowest frequency range the data, a~ter transformation
into the mean-wind cooordinates, are subjected to an 80-point non-over
lapping averaging for analysis in the frequency range between 0.00031-
1.25 Hz.
The last step of the data analysis is the spectral analysis of the
high, middle and low frequency data in the frequency range QfO~0244-
100 Hz, 0.00153-6.25 Hz and 0.00031-1.25 Hz respectively. Spectral
estimates are calculated for each block using a specially developed
Fast Fourier Transform algorithm [4]. The combined averaging
technique is employed, averaging first all the block estimates at a
given frequency (ensemble averaging) and then averaging these results
13
over appropriate frequency intervals (frequency averaging).
In total, 24 one-hour data records were generated with the hot
film system. Nine runs were generated during warm summer afternoons
of the year 1976. This set of data was acquired for southerly winds
only, and the detailed results are presented in Reference 3. The re
maining 15 runs were acquired during the spring of 1977 for winds
of northwesterly direction. For some of these data records, data
were acquired simultaneously with the cup-vane system.
The hot-film system is extremely delicate. Lightning and power1ine
fluctuations have often caused difficulties with the operation of the
system. It is very seldom that the entire system is fully operational
Clt one particular time. The hot-films also have a tendency to undergo
l:esistance shifts. If an appreciable shift is detected, the probes
are reca1ibrated. Corrections have to be made for changes in cable
resistance due to changes in ambient temperature. Similarly, heat
transfer corrections have to be made for changes in temperature.
Uecause of the uncertainties in these corrections and other variations,
the mean v!'!locity and mean temperature measured with this system
are not reliable. Resistance shifts in the active part of the system
(films Clnd cables) result in a parallel shift of the heat transfer
(voltagE~)/ve10city calibration C~rve. This shift of course will
affect the mean quantities a great deal but should not affect the
ealculated turbulence quantitites as much. As pointed out in
Referenee 5, the results from these instruments become less accurate
for wind directions of +40 0 and ±90° with respect to the probe
axis. Consequently, the probes are rotated in the direction of the mean
wind prior to data acquisition. For southerly winds this is no problem
14
since these winds especially at the higher elevations are very steady
and have low turbulence levels. For northwest winds the alignment
with the mean wind is more of a problem because of the presence of long
period fluctuations in direction. Precautions were made as much as
possible to ensure that data of the highest quality were acquired,
and it is believed that the measured turbulence quantities fall within
an accuracy level of less than 10%.
5. PLANETARY BOUNDARY LAYER (PBL)
The planetary boundary layer (PBL) may be defined as that part
of the atmosphere where the effect of the earth's surface is directly
felt. The flow structure of the boundary layer is extremely complex
due to the variability of surface roughness changes in terrain, changes
in surface temperature, variability of water vapor, presence of clouds
and the fact that the flow is turbulent. Consequently, a simple model
describing all the variables in the PBL such as velocity, wind direction,
temperature and humidity, and covering all possible conditions is still
not available.
The unstable or convective PBL is characterized by a strong upward
heat flux from the surface and by strong vertical mixing due to positive
buoyancy forces. Under these conditions above the surface layer, a
well-mixed layer exists with an almost uniform potential temperature
and an almost constant wind speed and direction. Due t·o the convective
mixing a relatively sharp inversion is created on top of the mixed layer
that delineates the depth of the PBL. Above this inversion the atmo
sphere is relatively undisturbed by the presence of the earth's surface
and only gravity waves are generated. During the course of a sunny day
15
the convective PBL increases in depth as the land surface heats up and
extends up to the inversion layer, which is the result of the convective
mixing of unstable air below with warmer stable air above. Under these
conditions the convective PBL can be characterized by three different:
layers, which are, starting from the earth's surface:
1) a surface layer, 2) a mixed layer, and 3) a capping inversion. The
existence of this model for the convective PBL is based on theoretical
consid,erations [6], numerical flow modeling [7,8,9,10], laboratory flow
modeling [11] and actual observations [12,13,14].
In the case a layer of stable air is present at the surface, and/or
one or more stable layers exist at higher altitudes, no capping inversion
exists. In stable air, turbulence is suppressed and may vanish completely
under extremely stable conditions, and the air in the different layers
becomes uncoupled as a result of reduced mechanical mixing. No simple
model is available to describe this situation. A surface-based inversion
with or without one or more stable layers at higher elevations has been
observed over water surfaces, as warm air flows over the cooler water
during either day or night [3,18]. Under these conditions a low-level
wind maximum (surface jet) is usually observed and the interpretation of
the wind fluctuations near the surface is often complicated as a result
of the co-existence of turbulence and internal gravity waves. These two
phenomena have quite different properties although a non-linear inter
action.. may exist between them.
Mlore complications are introduced in the flow analysis of the PBL
when c.louds are present and condensation of water vapor occurs within
the PBL. As condensation takes place in the layer where the clouds are
lO'cate!d, latent heat is released and an inversion layer develops below
16
the cloud base that is responsible for some degree of suppression of
the turbulence. Under these conditions the large-scale plume structure,
normally present in the convective PBL, does not develop resulting in
a considerable suppression and reduction in low-frequency velocity
fluctuations. This phenomenon in turn results in reduced velocity
variances and an increased roll-off in the low-frequency range of the
spectra of the horizontal velocity components.
Just above the earth's surface the shear-production of turbulence
dominates the buoyant production or suppression of turbulence. However
the importance of the shear-produced turbulence diminishes with height
as the effect of buoyancy on the turbulence gains importance. In the
layer just above the earth's surface where shear-produced turbulence
dominates and the Reynolds stress is nearly constant, the velocity is
adequately expressed by the well-known logarithmic law, provided the
roughness of the terrain is uniform. As buoyant production or buoyant
suppression of turbulence gains importance with height relative to shear
production, modification of the logarithmic velocity profile will result.
The buoyant production of turbulence is associated with the upward heat
flux of sensible and latent heat, which usually occurs when the atmosphere
de near the earth's surface is unstably stratified (dz < 0). The buoyant
suppression of turbulence is associated with the downward heat flux
de which usually occurs when the atmosphere is stably stratified (dz > 0),
where e is the local potential temperature.
Under strong-wind conditions on a sunny day, the layer in which the
shear-produced turbulence dominates increases in height, and the logarith-
mic velocity profile exists to higher elevations before it is eventually
17
modified as a result of convective activity, even under the strongest
wind conditions.
Following this discussion of the PBL, it is unlikely that its
flow characteristics are based exclusively on mechanically or shear-
produced turbulence. Experimental results describing the mean flow
and turbulence in the PBL, under a variety of stability conditions
ranging from unstable via near-neutral to stable conditions, clearly
indicate that some basic d:lfferences exist between the PBL and the
typical zero-pressure gradient wind-tunnel boundary layer. The most
noticeable differences are observed in the evolving convective PBL
on a sunny day with the mixed layer occupy:ing about 90% of the height
and the surface layer consisting of the lower 10% of the PBL. In the
mixed layer the velocity and wind direction are practically uniform
[12,13]. Also, the downward entrainment of heat and momentum at the
base of the capping inversion is a phenomenon that makes the assumption
of vanishing turbulence at the edge of the boundary layer untenable
[12,13,14).
The entrainment of heat and momentum of different magnitude and
direction into the mixed layer is mainly responsible for the low
frequency fluctuations observed in the horizontal velocity components
down to near ground level [19,20]. These low-frequency fluctuations
are not present under stable conditions, although as mentioned before,
low-frequency fluctuations associated with gravity waves have been
observed under stable conditions [15,18).
An abrupt discontinuity (increase) in low-frequency content has
been observed in the spectra of the horizontal velocity components as
the thermal stratification changes from stable to unstable [19,20].
18
Similarly, the variances of the u and v velocity components and the
reduced peak frequencies of these spectra increase abruptly as the
stratification changes from stable to unstable. These abrupt
changes have been observed in the horizontal velocity components
u and v only, and do not occur in the vertical velocity component w.
Experimental evidence of the abrupt increase in variance of the u
velocity component and of the total velocity variance as the flow
changes from slightly stable to slightly unstable is clearly presented
by Busch [21]. For slightly stable stratifications the ratio of the
standard deviations of the vertical and longitudinal velocities
a la = 0.36 and for slightly unstable stratifications this ratio w u
is reduced to 0.25.
It has been observed that shortly before sunset the convective
boundary layer disintegrates suddenly in a matter, of minutes as a
surface-based inversion begins to develop [22,23]. During the
night (or when there is a heavy cloud cover and water-vapor condensation
in the PBL) the large-scale turbulence structure of the convective PBL
is suppressed and no low-frequency components are present in the u and
v velocity spectra near the earth's surface.
Some of the experimental evidence, used previously in support of
the flow description of the PBL, was obtained under relatively
strong wind conditions. For example, during the Minnesota experiment
[12] two data records (2A1 and 2A2) were acquired each with an hourly
mean velocity of approximately 10 mls at a height of 10 m. For these
two data records the mean velocity is practically uniform above 60 m,
which is typical for the mixed layer. Also the standard deviations
19
of the three velocity components do not change appreciably with elevation
up to an elevation of 1220 m, indicating that there is little evidence
of vanishing turbulence at the edge of the PBL. In addition, the
horizontal turbulence stresses uw and vw exhibit minima near the
surface and a considerable upward heat flux is observed up to the top
of the mixed layer where the heat flux changes direction due to entrain-
ment of warm air at the base of the capping inversion. The gradient-
Richardson numbers for these two records of Reference 12 are approximately
-0.30 and -0.46 respectively, which excludes these records from the near-
neutral stability category.
In. another example, extremely strong winds ·with mean velocities up to
28 mls have been observed at 10 m heights over water [24]. The turbulence
intensi.ties of the u and v components show an abrupt increase for mean
velocities in excess of 12 mls (0" Iu from G% to 15%, 0" lu from 8% to u v
11.5% and 0" lu remains constant at 5%). There is strong evidence that w
helical vortices, which form over the ocean under near-neutral conditions,
are responsible for this increase in turbulence intensity of the
horizontal velocity components. Also nocturnal jets near the earth's
sUrfaCE! with a jet velocity of approximately 20 mls and a gradient
Richardson number of +0.5 have. been observed in Nebraska [16].
Based on these examples and other experimental evidence presented
in the cited references, there is no basis for the assumptions that
under strong wind conditions t.he turbulence in the PBL is purely of
mechanical origin and that the PBL flow-structure is similar to that
of a ZE!rO-pressure gradient wi.nd-tunne1 boundary layer with vanishing
turbulEmce at the free stream.
20
For the benefit of structural engineers and wind engineers who
deal with a large number of problems requiring the knowledge of the
mean flow and turbulence in the atmosphere, the atmospheric flow
characteristics near the earth's surface under strong wind conditions
have been reviewed and summarized in several review papers [References 25
through 31]. The input data to most of these review papers were obtained
from many different sources and include multi-level tower data and airplane
data usually taken over horizontal terrain with near-uniform roughness.
The results from these strong-wind experiments have formed the basis
for empirical models of the wind structure near the earth's surface
and of the atmospheric flows at hi~her elevations. In many of the
review papers it is assumed that under strong wind conditions the
turbulence is purely mechanically generated, and that buoyant produc-
tion and suppression of the turbulence can be neglected. Under these
conditions it is then assumed that the atmosphere is neutrally stable
and the mean and turbulent flow behave similar to the flow in a
zero-pressure gradient turbulent boundary layer developed in a low-speed
wind tunnel.
The review papers generally indicate tht the PBL under strong wind
conditions is neutrally stable but may be modified as the result of
thermal effects. However, it is generally assumed that under strong
wind conditions there is sufficient mixing that thermal effects can be
completely ignored. The neutrally stable PBL-flow model, which is
generally assumed in the review papers, has never been experimentally
verified at different sites for a large variety of strong wind con
ditions. Nevertheless, it is widely assumed by many engineers and
21
scientists in the field on wind engineering that this model provides
an adequate description of the flow in the PBL under strong wind
conditions.
Upon reading of the review papers one does not always get a clear
picture when neutral conditions exist. Deaves and Harris [30] do
not give any conditions for the existence of neutrally stable flow
in the PBL other then the mean velocity has to be strong enough.
Counihan [26] considers only those data which specifically indicate
adiabatic or near adiabat.ic conditions, or which have wind speeds in
excess of 5 mls at a height of z = 10 m. ESDU [27,28,29] considers
neutral stability to exist when the hourly mean wind at a height of
10 m is greater than 10 m/s. Using the gradient Richardson number
as a measure for the degree of thermal stability in the PBL, near
neutral conditions are suggested to exist by Teunissen [25] for IRil <O:~.03
and by Panofsky [31] for IRil<O.Ol for flow near the surface.
Panofsky [31] also states that wind shear, and therefore mechanical
produlCtion of turbulence, decreases rapidly with height and that the
effect of buoyancy becomes progressively more important. Consequently
thermal effects can no longer be neglected even under strong wind
conditions for heights above approximately 50 m.
Most of the observations at Wallops Island presented in this
report have been made with the mean wind speed at the 76 m (250 ft)
level between 10 m/s and 20 m/s and therefore can be considered to be
long to the strong-wind cat.egory. However the results of these ob
servations will be presented and discussed in the framework of non
neutral PBL-flow.
22
6. RESULTS AND DISCUSSION
This chapter deals with the discussion of the mean flow and
turbulence measurements obtained with the two instrumentation systems
mounted on the 76 m (250 ft) meteorological tower at Wallops Island,
Virginia.
Data were acquired with the cup-vane and resistance temperature
probes during a period of more than 4 years (August 1974 - December 1978).
Most of these data records were taken during daytime, although some
night records are included. The prevailing wind directions at this
sjra are southerly during the summer and fall, and between west and
north during winter and spring. Consequently, for these two direction
sectors many data records are available but the data base for on-shore
winds in the sector between northwest and south is limited. For this
set of data, mean velocity U, mean direction ~, mean temperature T,
the turbulence variances cr and cr *, and the turbulence integral scales u v
LX LZ, LZ were obtained for each data record, based on the measure
u' u v
ments from 5 levels at 15.3 m (50 ft) intervals on the 76 m (250 ft)
tower.
A second set of data was taken with the hot-film instrument
system. Nine data records, each one hour long and for southerly wind
directions, were acquired during 3 days in July and August of 1976 [3].
In addition 14 data records for winds from northwesterly directions
were acquired during several days in March, May and June of 1977 [23].
*The symbol cr refers to the standard deviation, but for simplicity the word variance will be used in the text.
23
For these data records the following parameters were obtained: variances
and covariances of the three turbulence components and temperature,
spectra and cospectra, and lon.gitudinal integral scales LX, LX and LX u v w'
all measured at the same levels as the cup-vane instruments and at 9.1 m
(30 ft). For some of these records data were acquired simultaneously
with the cup-vane system, allowing direct comparison of the measure-
ments,with the two instrument systems.
6.1 ME!an Wind and Mean Temperature Profiles
Fc,r near-neutral stratification the wind profiles over homogeneous
terrain obeys the relation
U =(U*/K) In(z/z ) o 0
where U* is the friction or shear velocity, which is ideally equal to o
lTo/p, where Lo is the surface stress and p the air density, K is
Von Kax~an's constant taken at 0.4 and z is the roughness length. o
The he:l.ght z below which the log-law (2) is valid depends on the
stabi1:1.ty of the flow or on the relative importance of the buoyant
(2)
production or suppression of the turbulence with respect to the mechani-
cally produced turbulence. The height where the measured velocity
profilE! starts to deviate from the logarithmic profile varies a great
deal and the deviation increases progressively for increasing heights.
Under strong convective conditions and under extreme stable conditions
the neaLr neutral part of the surface layer is well below 15.2 m (50 ft),
the hei.ght of the lowest mean-velocity measurement. Consequently under
these conditions it is incorrect to fit the log-law (2) to the measured
mean-ve.locity data in order to obtain values of z . and U*. o 0
24
The theoretical velocity profile over uniform terrain for the
upper part of the surface layer where non-neutral stratification
exists, is based on the Monin-Obukhov similarity theory and is given
by [32]
(3)
where ~ is a universal function of the stability parameter zit and
t is the Monin-Obukhov length defined (in the absence of moisture) as
[32]
c p e U*3 t =-L (4) K g Qo
where T is the absolute temperature, U* = IT7P and Q is the surface 0 0
heat flux, which can be approximated by c p we measured in the surface p
layer. The stability parameter, zit, depends on the gradient Richardson
number [32]
y + dT/dz Ri -.8. a ,
- T (dU/dz)2 (5)
where y is the adiabatic lapse rate (O.Ol°C/m for dry air and a
O.0065°C/m for air saturated with moisture). With the use of the
expressions relating Ri, zit and ~ as given by Panofsky [31], the
departure from the logarithmic profile due to buoyancy effects
in the surface layer can be obtained so that (3) can be used to
obtain estimates of z and U*. o
This approach can be taken. for velocity profiles measured in
the surface layer, which is loosely defined as that part of the PBt
where the horizontal stress and vertical heat flux are nearly con-
stant. The height of the surface layer is frequently estimated as
25
the 10wE~r 10% of the convective PBL depth, Z., and is limited to a ]-
height z<L [13].
For stable conditions L varies a great deal and is relatively
small [JL7]. The surface layer is not well defined and often is only a
few metE!rS high [33]. Under these conditions the velocity profile for
Ilon-neutral conditions cannot be used for estimation of z and U*, with o 0
velocity measurements taken well above the surface layer.
In general, Monin-Obukhov (M-O) similarity theory can only be applied
:I.n the surface layer, when the flow conditions are stationary, when
surface conditions are uniform (roughness and temperature), and over
level tE~rrain without major topographical features such as mountains
E!tC. If the measured velocity data are linear with In z, then this
observation does not automatically guarantee that the profile is truly
logaritt®ic. For the majority of sites including the Wallops Island
site the~ above conditions are not met, and application of the M..,O
similarity concepts such as the empirical flux-profile relationships,
which are based on data with the above conditions satisfied, is
questionable and should probably be avoided.
For neutral conditions the temperature profile should follow
the adiabatic lapse rate of either 0.01 or 0.0065°C/m for dry and
saturated air respectively. It is very seldom, even under the strongest
wind conditions, that a truly neutral stratification is encountered
for any length of time. The usual daytime thermal stratification near
dT dT the surface is - - > y (unstable stratification) and - -d < Y (stable dz a z a
stratification) for nighttime. There is a short time around sunrise
and in the late afternoon before sunset when a neutrally stable strati--
fication is observed near the surface in the first 10 or 20 m. The
26
stratification of the air at higher elevations varies a great deal with
time of day, cloud cover and season. For southerly flows over the ocean
the air temperature is usually higher than the water temperature and
a surface-based inversion (-dT/dz < y ) exists day and night. a
Changes in temperature and velocity profile occur also as the flow
experiences a sudden change in surface roughness and surface temperature,
and an internal boundary layer (IBL) develops as the flow adjusts itself
to the new surface conditions. On-shore winds at Wallops Island usually
experience an increase in roughness and an increase in surface temperature
in daytime as they cross the beach. For westerly winds the surface
temperature changes as the wind moves from the wet marsh over the warm
land surface during daytime. Within the observation height of 76m at the
Wallops Island site, temperature gradients vary greatly with height and
even change from positive to negative or vice versa. Under these condi~
tions the local Richardson number is not a true indicator of the overall
thermal stratification of the observed flow and ~O similarity does not
apply.
Typical temperature profiles for southerly winds on a summer afternoon
are shown in Figure 5. These temperature profiles clearly indicate the
stable conditions above the IBL in the early afternoon, and the surface
cooling in the late afternoon. Daytime and nighttime temperature profiles
for strong westerly winds are shown in Figure 6. During daytime the
surface temperature is maximum and gradually decreases with height. On
a rainy day with low cloud cover a stable stratification is observed above
40 m. Neutral temperature profiles throughout the tower height are seldom
observed; the second temperature profile shown in Figure 6 is the closest
to an observed neutral temperature" profile. Typical nighttime temperature
profiles show a stable stratification near the surface with near neutral
27
conditions above 15 m.
Ratios of mean velocities at the 15.2 m (50 ft), 30.5 m (100 ft),
45.7 m (150 ft) and 61 m (200 ft) levels with respect to the mean velocity
at the 76.2 m (250 ft) level are shown in Figure 7. The plotted data in
this figure represent the mean and the plus and minus standard deviation
from the mean, as well as the maximum and minimum values of the velocity
ratios for the data records acquired in each IS-degree sector. Because
the mean velocity at higher elevations is less disturbed by local surface
obstacles and is outside the developing IBL, the mean velocity at the
76.2 m (250 ft) level was chosen as the reference velocity. For the one
sector between 0° and 15° and the 5 sectors between 90° and 165° (east
southeast) the number of data records available in each sector was four
or less, an insufficent number for determination of the standard deviation,
and only the average velocity ratios are presented.
B4~tween 230° and 360° (southwest-north) the variation in mean wind
speed between the different elevations is relatively minor under moderately
strong wind conditions. Conversely, for southerly winds in the sector
160° < ~ < 230, large variations in mean wind speeds relative to the
mean velocity at the 76.2 m (250 ft) level are observed. For example
iIi the sector between 180° a.nd 195° the mean wind-speed ratio, V50/V250,
varies from 0.4 to 0.8, clea.rly indicating the great variability of mean
velocity distribution near the surface. Relatively large variations of the
mean velocity ratios are also observed for wind directions approximately
parallel to the island for sectors 30°-45° and 210°-225°. For winds from
these two sectors a small change in wind direction gives rise to a large
variation in upstream terrain roughness. Generally the variations of
the mean velocity distribution near the surface for on-shore winds are
28
much larger than for westerly winds. For on-shore winds, the changing
development length of the IBL, the thermal stratification and other
atmospheric conditions have a great effect on the velocity distribution
near the surface as the data indicate. For southerly winds, a surface-
based inversion exists during the daytime as the result of warmer air flowing
over the colder water [3]. Under these conditions the existence of a sur-
face jet with a maximum mean velocity (jet velocity) within a few hundred
meters from the surface has been observed. The jet velocity has been
observed as low as 45.7 m (150 ft) as shown by the maximum values of
the velocity ratio V150/V250 of larger than one for wind directions be-
tween 170° and 210°, and by one of the velocity profiles of Figure 8.
Typical strong-wind velocity profiles for wind directions between
southwest and north are shown in Figure 9. The stronger wind profile
is linear over the entire observation height when presented in
semi-logarithmic coordinates. In the second profile with somewhat lower
velocities, a distinct "kink" is observed. The majority of the measured
velocity profiles for westerly wind directions shows similar "kinks".
The values of z obtained from the velocity profiles above the "kink" o
are much too small for the upstream terrain. Moreover the two velocity
profiles of Figure 9 are approximately from the same direction. Therefore
the "kinks" in the profiles cannot be the result of an upstream change
in roughness, but must be interpreted as the beginning of the transition
from the surface-layer flow to the uniform mixed-layer flow [13].
6.2 Velocity Profile Parameters z and a o
The roughness length, zo' can be evaluated from fitting of either
the logarithmic law (2) or the non-neutral profile law (3) to the
measured mean velocities in the surface layer. The height of the
29
surface layer in which (2) and (3) are valid is roughly defined a,s the
layer near the surface which t.he fluxes uw and we are approximat~ly
constant and not less than 80% of the surface stress~ ~p~ and surfaee o
heat flux, Q /c p, respectively. Measurement of the turbulent fluxes at o p
WallQPs Island indicate that t.he thickness of the surface layer varies a
great deal with wind speed, wind direction and thermal strat:l.fication.
For on-shore winds the top of the surface layer is generally
below the lowest observation level of 9.1 m (30 ft) and for the
strongest westerly winds the surface layer extends above the height of
the tower. For on-shore winds when an IBL develops as the air crosses
the beach, turbulent heat-flux measurements show an upward flux at the
lower levels and a downward flux at the higher elevations [3]. Transi--
tion usually occurs between the 15.2 ~ (50 ft) and 30.5 m (100 ft) levels,
which should correspond to the height of the IBL at the tower location.
The flow above the IBL is still associated with the ocean surface, but is
w~dl apove the surface layer, which for southerly winds extends only a
fl~W meters above the water surface. Consequently no values of the
r()Uglmess length, zo' for either the land or ocean surface can be ob
tained f:r:-om the measured velocity profiles. Despite the fact that two
velocity vrQ£iles of Figure 8 show a linear variation of velocity with
Inz, fitting of the log-law (2) to the velocity data leads to values
of Zo of the order of Im~ which are much too high for either the exist:lng
4pst~eam terrain or the ocean. The parameters describing the on-shore
turbulen"1: flow depend in part on the roughness of the underlying surfaee
(ocean) but also to a great extent on the stability of the flow. Con-
sequently, knowledge of the roughness length, zo' for flows over the
oeean is not sufficient for the prediction of the turbulence
30
parameters for higher elevations as presented by ESDU [28,34].
For westerly winds the local shear velocity U*=[uw2 + vw2]1/4
is approximately constant over most of the observation height, in-
dicating that the surface layer extends well above the lowest obser-
vation level of 9.1 m (30 ft). Also the values of the local shear
velocity, U*, compare quite well with the profile friction velocity
U~, obtained from mean velocity measurements below the profile "kink".
The roughness length, z , obtained from fitting of the log-law (2) to o
the measured mean velocities below the profile "kink", vary with up-
stream roughness and thermal stability of the flow in the surface layer.
Figure 10 shows the variation of z with wind direction, clearly o
indicating the effect of the 6.5 m high storage bunker for mean wind
directions of 270°. In the sector between 350° and 30° the upstream
terrain is much rougher which results in higher values of z. For the o
sectors between 230 0 and 260°, and 280° and 340°, the terrain is
reasonably uniform and the average value of the roughness l~ngth,
z =0.034 m , corresponds quite well to the predicted values .of o
ESDU [28,29] for similar terrain. In Figure 11 all values for z ob-0'
tained for wind directions in the before-mentioned sectors, are shown
as a function of the gradient Richardson number evaluated at z=15.2 m
The values of z were obtained by fitting of the log-law to o
the velocity measurements below the profile "kink". The scatter of the
data is appreciable for Ri15>-0.15, however the values of Zo decrease
rapidly for Ri15<-0.15. At this point the mean velocity
profile in the surface layer can no longer be estimated with the log-
law (2), as stability effects start to dominate. An attempt was made to
estimate the roughness length by fitting the non-neutral profile (3)
to the mean velocity data above the "kink" as outlined by Panofsky [31].
31
However, this correction technique leads to unrealistic estimates of
z , which seem to indicate that the velocity profile above the "kink" o
is not part of the surface-layer profile but instead is the transition
of the surface layer profile to the mixed layer profile.
In Table I mean profile parameters and turbulence parameters are
presented for strong-wind data records with wind directions in the
above-mentioned sectors with near-uniform upstream terrain and measured
with the cup-vane system and the hot-film anemometers. For the data
records with a thermal stratification close to neutral the semi10garith-
mic velocity profiles (S) are generally linear throughout, and for the
records for which the flow is more unstable, kinks (K) start to appear
in the~ profile.
Engineers generally favor the power law as an expression for the
velocity profile through the entire PBL over uniform terrain
(6)
Although under certain conditions this law has fitted observed profiles
over uniform terrain quite well, it is now generally accepted that the
log-law (2) is preferable over the power law (6) for flow in the sur-
face layer. Also the power law (6) cannot be expected to be a good
apprOldmation for wind profiles 'in the convective PBL. The latter seems
to be prevalent over the North American continent even under the strongest
wind (~ond it ions.
The mean velocity profiles of records 2A1 and 2A2 acquired during
the Mjlnnesota 1973 atmospheric boundary layer experiment [12] as well as
the v,docity profile acquired at the Boulder Atmospheric Observatory
(300 In mast [35]) on the morning of the 11th of September 1978 under
32
extremely strong wind conditions (20-minute average velocity of 18.67 mls
at z=lO m) show clearly that the power law with one constant exponent
does not fit the velocity measurements. The measured strong-wind
velocity profile at the Boulder site starts to deviate from the log-law
(2) at z=22m a.nd becomes more or less uniform above 200 m.
On the other hand mean-velocity measurements made at 3 levels (10m,
80 m and 200 m) of the 2l3-m mast at Cabauw, the Netherlands, under
extremely strong westerly winds (30-minute average velocity of 22.2 mls
at z =10 m) [36] indicate that the power law (6) fits these data quite
well. The Cabauw tower is operated by the Royal Netherlands Meteorological
Institute and is located about 50 km east from the North Sea coastline.
For westerly winds the upstream terrain is flat low-lying pastureland with
a very shallow water table. Simultaneous temperature measurements
made at 8 levels between 2 m and 200 m indicate a near- neutral thermal
stratification with possibly a minimal upward heat flux from the mostly
wet upstream terrain that inhibits the formation of a convective PBL,
which is typical for the observations of the two North American sites.
As Panofsky [31] has pointed out the power law can be expected
to fit the velocity data in the surface layer only over a limited height
range. In the near-neutral surface layer where the logarithmic law
applies the power-law exponent can be approximated by [31]
(7)
clearly showing the variation of a with roughness, z , and geometric mean o
height, Izl z2 where zl and z2 represent the elevation boundaries of
the layer over which a near constant a may be expected. Near the sur-
face a has a relative large value depending on the value of z , but o
33
decreases with height and should approach zero outside the surface
layer as transition to the mlxed layer takes place.
For westerly winds at the Wallops Island site, power law exponents
based on the mean velocities of the three lower levels (zl=15.2 m,
z2=45 .. 7 m) are shown in Figure 12 as a function of the roughness length,
z. Similarly, power law exponents based on the mean velocity measureo
ments under strong-wind conditions (no "kinks") at all five levels and
the three highest levels are shown versus roughness length, z· in 0'
Figurl~s 13 and 14 respectively. Panofsky's relation (7) fits the
measurements extremely well except for the results based on the mean
velocities of the three highest levels (Figure 14). The reason for this
is that on a semi-10garithmie plot of U vs. In z, the results of the
higher levels fall close together and may seem to vary in a linear
fashion. However in reality the data already deviate from the log-law (2)
and all three prediction methods, which are based on known roughness
lengths, overestimate the power law exponents obtained from the measured
veloc:ity profiles. No attempt was made to obtain power law exponents
for the velocity profiles of on-shore winds, because of the variability
of the profiles, the presence of an IBL and the absence of values for
z • o
6.3 Turbulence Intensities
Average turbulence intensities of the horizontal velocity components
u and v measured with the cup-vane instruments at 5 levels, are shown in
Figures 15 and 16 for each 1.5-degree sector. In addition to the average
turbulence intensities, the maximum and the minimum observed turbulence
intensities for each sector and the mean intensity plus and minus the
34
standard deviation are shown. In the sector 0°-15° and the 5 sectors
between 90° and 165° only average turbulence intensities are shown,
because the number of data records available in each of these sectors
is four or less. For on-shore winds a Iu is approximately 10% at u
z=15.2 m decreasing to 5% or less at the highest observation level
(z=76.2 m). For westerly winds the a Iu decreases from about 20% at u
the lowest level to about 14% at the highest level (Fig. 17). In
general the variation of the turbulence intensity, a IU, in each u
section is relatively small, except for the two sectors between 195°
The turbulence intensity of the v component, a lu, shows much the same v
pattern, although some differences can be observed. For on-shore winds
the average value of a lu at the lowest observation level (in the IBL) v
is about 8%, decreasing to about 4% at the highest level (above the IBL)~
For southerly winds in the three sectors between 165° and 210° turbulence
intensities for both u and v components of less than 2% have been ob-
served at the highest observation levels above the IBL. These observations
have been made under stable conditions, and on several occasions the
mechanical turbulence has been observed to vanish completely at these
elevations at mean velocities of 10 mls and higher.
The average turbulence intensity alU is nearly constant at each u
level for wind directions between 240° and 345°. The ~verage turbulence
intensity a lu at each level increases with wind direction from 240° to v
about 300° where a maximum is reached and decreases gradually with wind
directions from 300° to about 50°. Figure 17 shows the variation of the
average turbulence intensities of the u and v components with height
in the two sectors 240°-255° and 300°-315°. The averaged results are
also compared with the estimates of Teunissen [25] and ESDU [28] both
35
based on the roughness length, z , whose average is 0.037 m for either o
one of these sectors [Fig. 10]. Although the upstream terrain for the
two sectors is about identical, the values of cr Iu in the sector bev
tween 300° and 315° are considerably higher than those for the sector
between 240° and 255°. The average values of cr Iu for the latter sector v
correspond extremely well with the Teunissen estimate, but for the
sector between 300° and 315° the average values of cr lu are about 3 v
to 4% higher. On the other hand the average values of cr lu for these u
two sectors are about identical and fall between the estimates of
Teuriissen [25] and ESDU [28].
A possible explanation for this unusual behavior of cr Iu can be v
derived from the visual inspection of several years of strip-chart re-
cordings of wind speed and direction obtained continous1y from a
propeller-vane anemometer located about. 1 km northeast from the tower
on Wallops Island. These recordings clearly show that for northwesterly
winds the instantaneous wind direction experiences frequently large
f1uctua.tions toward the north (Fig. 18). These direction fluctuations
are larger than usual and are not normally distributed as shown in
(Fig. 19), which explains the large values of both (J Iu and the cov
variance vw measured in this sector.
The turbulence intensity, (J lu, which is constant over near-uniform u
terrain for westerly wind directions (Fig. 17), instead varies with time
of day. Nighttime measurements are systematically 1% higher than the
daytime~ results (Fig. 20). The vertical distribution of the vertical
turbu1emce intensity, cr lu, based on five strong-wind data records obw
tainedwith the hot-film system, is shown in comparison with the ESDU
[28] and Teunissen estimates in Figure 21.
36
The vertical distributions of the turbulence intensity of all
three components for southerly winds are shown in Figures 22, 23 and
24. For these wind directions an internal boundary layer (IBL) develops
as the air crosses the beach. Based on the change in direction of the
daytime vertical heat flux we, which is expected to be positive in the
IBL and negative above it [3], the height of the IBL at the tower must be
between the 15.2 m (50 ft) level and 30.5 m (100 ft) level. This
observation agrees with the relationship given by Elliot [37], which
predicts a height of the IBL of approximately 26 m, based on an upwind
roughness length over the ocean of z 1=0.001 m and a downwind roughness o
length of z "=0.01 m and a development length of 300 m. The average o
turbulence intensities of the uand v components from nine data records,
obtained with the hot-film instrumentation during sunnner afternoons, com-
pare surprisingly well with the average turbulence intensities from 29
data records, obtained during winter and early spring under strong wind
conditions with the cup-vane system (Figs. 22,23,24). The turbulence
intensities in the IBL compare reasonably well with the ESDU [28] pre-
dictions based on a roughness length of z =0.01 m. However above the o
IBL the ESDU [28] predicted values of the turbulence intensities, based
on a roughness length z =0.001 m, overestimate the actual measured values o
by a factor of two. The turbulence intensites for run 7 [3] obtained
under extremely stable conditions (Rf =+25.5, U=10.7 mls atz=6l m), which
are included in Figures 22 through 24, show lower than average values above
the iBL. Under similar conditions it has been observed that the turbulence
vanishes completely for some time. The variation of the ratios of
average turbulence intensities, a la and a fa , for all wind directions v u w u
is shown in Figure 25. The measurements indicate that the values of
37
these ratios are more or less independent of height, and the data
shown in Figure 25 are not only the ratios of the average turbulence
intensities of all data records in each sector but are also averaged
over all observation levels. For most wind directions the values of
a /a are between 0.75 and 0.85 in comparison with the predictions v u
of 0.75 and 0.80 by Counihan [26] and Teunissen [25]. The estimates
of ESDU [28] are not single valued but are dependent on height and rough-
ness length. Smaller values than 0.75 for a /a are observed for wind v u
directions approximately parallel to the island at ~~40° and ~~230o.
Values of a /a higher than 0.85 are observed for southerly winds for v u
~~170° and for northwesterly winds in the sector 280o<~<350o. That
the values of a /a are relatively high in this last sector is no v u
big surprise since often large direction fluctuations have been
observed in this sector (Fig. 18). Values of a /a and a /a obtained v u w u
from the hot-film data records are also sh()wn in Figure 25. The val.ues
of a /a obtained with this system compare quite well with the cup-·vane v u
results. The values of the ratio, a /a , for winds from the sector w u
between south and southwest: I:!-nd northwesterly winds fall between 0.55
and 0.60 as compared to values of 0.50 and 0.52 as predicted by
Counihan [26] and Teunissen[25] respectively.
The results discussed so far in this report show clearly that at
the Wallops site large variations in mean as well as turbulent flow
occur varying with wind direction. The important observed deviations
from the simple neutral boundary-layer models are:
1. The development of an internal boundary layer (IBL) for on~shore winds as they cross the beach.
2. The existence of a surface jet for southerly winds with extremely low turbulence intensities (2% or less and occasionally
38
vanishing under stable conditions) coexisting with gravity waves.
3. Large direction fluctuations towards the north for northwesterly winds, observed during the daytime, which are responsible for higher than usual lateral turbulence intensities.
These observations were made under strong wind conditions and there is
no reason to believe that for still stronger wind speeds (maximum observed
speeds) these deviations from the neutrally stable boundary-layer model
will suddenly vanish.
6.4 Turbulence Integral Scales
6.4.1 x z z Integral Scales, L , Land L from cup-vane data u u v
In this section the distribution of the turbulence integral scales
obtained from measurements with the cup-vane system will
be discussed and compared with the estimates from several review papers
[25,26,28,29]. The streamwise turbulence integral scale of the u
x component, L , is calculated from the autocorrelation function (1), u
R (T), assuming that Taylor's hypothesis is valid. Averages of all u
x scales, L , obtained from the data records in each I5-degree sector, u
are plotted along with the maximum and minimum and plus and minus the
standard deviation from the average value in Figure 26 for each ob-
servation level. Averages only are plotted in the sector 0°_15° and
between 90° and 160° because of the limited number of available data
records in these sectors. The integral scales, LX, increase u
systematically with height and show a great deal of variability for all
wind directions. In general the magnitude of these integral scales
is larger for westerly winds than for on-shore winds except for southerly
wind directions. At the highest observation level two distinct extremes
for the maxima can be observed in the two sectors between 180° and 210°,
and between 300° and 330°. For southerly winds between 180° and 210°
39
the air is frequently stably stratified which under certain conditions
may lea.d to the coexistence of turbulence and low-frequency gravity
waves. If gravity waves are present the integral scale obtained from
the autocorrelation function can be expected to be high [3], a maximum
value of 800 m at the 76.2 m (250 ft) level has been observed. On the
other hand, if the low-frequency gravity waves are absent and only tur-
bulencE! of less than 5% intensity is present (Fig. 15), minimum turbu-
lence i.ntegral scales of less than 100 m have been observed for the
southeI'ly winds. In the sector between 3000 and 3300 large variations
x in L are also present, which are the result of either the presence u
or absE!nce of low-frequency velocity fluctuations as can be observed
from mE!asured u-spectra.
VE!rtical integral scales of the horizontal velocity components
obtainE!d from the measurements with the cup-vane system can be cal-
culated by integration of the vertical correlation coefficients of either
the u ()r v components.
00
J o
where
and i •• u or v
RZ (z')dz' ii.
1 T
T J ui(t,z) ui(t,z+z')dt
o
(8)
(9)
R~i(z') is the vertical correlation coefficient of either the u or v
velocity fluctuations measdured at two different levels separated by
a distance z'. For this research program with the Wallops Island
tower the separation distance, z', can be either 0, 15.2 m (50 ft),
40
30.5 m (100 ft), 45.7 m (150 ft) and 61 m (200 ft). The integration
(8) should be performed to the point where the correlation coefficient
R z(z') changes for the first time from positive to negative. However uu
only a maximum of five values of the correlation coefficients are
available for either upward or downward integration according to ex-
pression 8, and often the correlation coefficient has still a large
positive value for the maximum separation distance z'.
In order to arrive at a reasonable estimate of the vertical scales,
it is assumed that the vertical correlation coefficients of the hori-
zontal velocity components u and v decay exponentially in the same
manner as has been observed by Dryden et al. [38] for high Reynolds-
number grid turbulence, according to
z z R,.(z') = exp[-z'/L.] 11 1
(10)
Vertical integral scales of either the u or v velocity components can
then be obtained from a least-squares fit of (10) to the available
measured correlation coefficients.
Because of the non-symmetric flow in the boundary layer, integral
scales obtained from upward and downward integration of the correlation
coefficients with the origin at a common point cannot be expected
to be the same. Instead a slightly different definition for the
vertical integral scale is used as· suggested by ESDU [29], where the
correlation coefficient is defined as
where i = u or v.
1 T
T
I v.(t,z+z')v.(t,z-z')dt 1 1
o
(11)
41
With thE~ use of this definition only one vertical integral scale is
defined for each height z. However this definition can be used for
the evaluation of the integral scales of the u or v component at the
45.7 m (150 ft) level only.
The distribution of the vertical turbulence integral scales of
the horizontal velocity fluctuations and obtained with the cup-vane
system is shown in Figure 27 and 28 respectively. Calculation of
these seales is based on expression (10). Integral scales obtained
from upward integration of the correlation coefficients are shown
as LUZt or LVZt on the figures and as LZt u
z or L t in the text. v For
the intE~gral scales obtained from downward integration the direction
of the arrow is reversed. The two-sided integral scales obtained
according to the definition (11) are shown as LUzl or LVzl on the
figures and as L~! or L~! in the text. Averages of all the integral
scales for each sector as well as the maximum and minimum, and the
mean plus and minus the standard deviations for each each sector are
shown in these figures. In the sectors 0°-15° and between 90° and
165° not: enough data records were available to calculate a' standard
deviation and only average values are plotted.
The integral scales LZ for on-shore winds are generally smaller u
than those for westerly winds by less than a factor of two. The
i.ntegral scale LZt measured at the 15.2 m (50 ft) level and the scale u
L~+ measured at the 76.2 (250 ft) level are of the same magnitude.
The smallest observed value of LZ is about 10 m for southerly winds u
without the presence of gravity waves. . Z The larger values of L are
u
also observed for southerly winds when gravity waves are present.
42
Large values of LZ are also observed for northwesterly wind directions u
between 300° and 320°. The upward integrated scales, LZ+ , obtained u
from the lower three levels for wind directions between 0° and 100°
Z show much less variation than the downward integrated scales, L + , u
from the upper three levels in the same sector. Comparison of the
three different scales, LZ+ LZ+ and Luzt of the measurements at u ' u +
the 45.7 m (150 ft) level shows that for winds between 0° and 100°
the upward and downward integrated scales are of the same magnitude
(37 m) with the scale L~t having an average value of approximately
43 m. For westerly winds between 250° and 360°, LZ+ is about 20 m u Z larger than L + • u
Z The vertical integral scale of the v-component, L , behaves in a v
Z similar fashion as the scale L , but the latter is generally u
twice as large. The magnitude of LZ for winds from westerly directions v
is about twice the value of LZ for on-shore winds.. Minimum values of v
just a few meters are observed for on-shore winds. Maximum values of
these integral scales are associated with winds from northwesterly
directions (~~3100) for which large direction fluctuations have been
observed (Fig. 18). For southerly winds (~~1800) no large maximum
values for LZ have been observed as for LZ. For on-shore winds the v u
magnitude of the three different scales at the 45.7 m (150 ft) level
are about the same but for westerly winds LZ+ > LZ+ with the value v v
of L~ t in between.
Figures 29 and 30 show the variation of the relative magnitude
x Z of the average turbulence integral scales L /L with wind direction u u
for each observation level. For southerly winds between 135° and 225°,
the ratio LX/Lz increases with height from about 3.7 at the lowest u u
43
level to values between 10 and 15 at the three highest observation
levels. For all wind directions outside this sector the ratio LX/Lz
u u
is 3.3 and 3.7 at the 15.3 m (50 ft) level and the 76.2 m (250 ft)
level respectively. The magnitude of the ratio of average integral
Z Z scales, L IL , is 1.9, generally independent of height and wind u v
direction.
6.4.2 Comparison of the Cup-Vane Scales with Predicted Values from References 25, 26' and 29
From the above discussion it is clear that the measured turbu-
1 . 1 1 LX LZ d 1z ., f . 1 . h . ence 1ntegra sca es , an vary s1gn1 1cant y W1t terra1n u u v '
roughness, wind direction, time of day, thermal stability and other
atmospheric conditions. In this section the measured integral scales
are compared with the predicted scales from either Teunissen [25],
Counihan [26] and ESDU [29].
In Figure 31 the averaged turbulence integral scales, LX obu'
tained from the cup-vane data are shown for two westerly wind-
direction sectors over near--identical terrain and are compared with
the estimates of References 25, 26 and 29. At the higher elevations
the values of LX in the sector 300 o <<P<3l5° are approximately 100 m u
larger than those in the sector 240 o <<P<255°. The Counihan [26] pre-
dictions match the measurements below 30 m but for higher elevations
all three references under-estimate the measured integral scales.
Averaged values of LZ and LZ for the same two wind direction sectors u v
are shown in Figures 32 and 33 and compared with predicted values from
Counihan [26] and ESDU [29]" The results show the decrease in magni--
tude when the change is made from upward to downward integration.
44
Values of ~ in the sector 3000<~<3l5° are about twice as large as those
in the sector 24Qo<~<255°, which is in contrast with the values of
LZ which are approximately identical in the two wind-direction sectors~ u
Z z The ESDU [29] predictions of both Lu and Lv match the measured data
in the wind sector 2400<~<255° reasonably well~ but underestimate
the scales for wind directions between 300 0 and 3150 for which large
wind-direction fluctuations have been observed [Fig. 18].
Figures 34, 35 and 36 show the averaged values of LX, LZ and u u
LZ obtained from the cup-vane data, in comparison with the predicted v
values for southerly winds. Since no measured values of roughness
length are available for this wind direction, values of Z were o
selected in accordance with the nature of the upstream terrain and
Table 1 of Reference 28. Below 20 m, for the flow in the IBL, a
roughness length Z =0.01 m is appropriate and for the flow above o
the IBL a roughness length of Z =0.001 m was selected. The measured o x longitudinal scales, L , are generally larger than the ESDU [29] u
Z Z predicted values, while the measured vertical scales Lu and Lv are
generally smaller than the ESDU [29] predictions. These results support
the likelihood that internal gravity waves exist in the surface-based
inversion over the ocean. The extent of the waves is much longer
in the direction of the flow than in the vertical direction because
of the suppression of vertical velocity fluctuations by buoyancy
forces in stable air. The results of Figure 30 show clearly the
large extent of
vertical scale,
the streamwise integral scale, L~, relative to the
LZ for southerly winds only. For these winds the u
scale ratio LX/Lz above the IBL varies between 10 and 15, for all u .u
45
other wind directions the magnitude of this ratio is of the order
of 4.
Longitudinal integral scales, LX, obtained for westerly winds u
over near-uniform terrain are dependent on time of day (Fig. 37).
The scales acquired during nighttime are approximately 50% longer
than those obtained from morning records with the scales from
afternoon data falling in between. The increase of the scales
during daytime can be explained with the increase of the height
of the mixed layer (See Fig. 5, Reference 13). However, no definite
explanation for the large integral scales observed during nightime
is available.
In Figures 38 and 39 individual values of the turbulence integral
scales, L~, at the 15.2 m (50 ft) and the 45.'7 m (150 ft) levels
a.re plotted versus roughness length, z. These results are for o
near-neutral strong wind data recorde for wind directions varying
between ~=250° and ~=30o. The scatter of the data is appreciable
as is to be expected since the previously discussed results also
in4ic~Fe variation of L~ with wind direction (Fig. 31) and time of
day (Fig. 37). The ESDU ~29] and Teunissen [25] predictions are
consistently lower than th~ measurements, while the Counihan predic-
tions [26] fit the measured integral s~ales reasonably well.
Simila:r:1y, th~ corresp<;mding vertical scales L~l and L~l are shown
versus r04ghness length, zo' :l..n Figures 40 and 41. The ESDU [29]
predicti.oq.~ :I;or L~l match the measured results quite well, while the
predict~.OIlS for L~l of the same source fall generally below the
measured data except in the range 0.1 m<z <1.0 m. o
46
6.4.3 The Direct and Spectral Methods for Obtaining Turbulence Integral Scales
Integral scales of turbulence are defined for any correlation
coefficient as the integral over the entire range of the independent
variable, which can be either time or space as previously discussed.
In practice the integration process is carried out between the origin
and the first zero-crossing. Time scales are related to the length
scales in the direction of the flow by assuming the existence of
Taylor's frozen turbulence hypothesis. The magnitude of the time scales
varies significantly depending on the presence of low-frequency fluctu-
ations and or trends. IIi order to omit these fluctuations from the
time-correlation coefficients, the data should be high-pass filtered
at some low frequency. The filtering of the cup-vane data takes place
as a result of the block averaging, excluding trends and low-frequency
fluctuations below 0.00195 Hz from the sample. Similarly, the hot-film
data are high-pass filtered at either 0.0244 Hz or at 0.00153 Hz de-
pending on whether the data are analyzed in the high-frequency range
(0.0244-100 Hz) or middle-frequency range (0.00153-6.25 Hz) (See section
4.2). Turbulence integral scales calculated in this way are said to
be obtained by the direct method.
An estimate of the size of the energy-containing eddies can also
be obtained from the Von Karman interpolation formula for the three-
dimensional power spectrum covering the wavenumber-range from the
energy containing eddies to the inertial subrange [39].
C(k/k )4 = ___ ..:;e_-::--::-=-
[l+(k/k )2]17/6 e
S(k/k ) e (12)
47
where k corresponds to wavenumbers in the range of the energy-conta:lning e
eddies. With the assumption of isotropic turbulence, expressions for
the physical realizable one-·d:lmensional spectrum functions Su (kl ) ,
Sv(kl ) and Sw(kl ) can be obtained (see expressions 3~72 and 3-73,
Hinze [40]). The wavenumber of the energy-containing eddies,
be replaced by the inverse of the turbulence integral scales,
k , e x Li
,
can
in the
appropriate spectrum functions and the constants can be adjusted to fit
the measured spectra. The one~dimensional Von Karman spectrum functions
obtained in this manner and presented by Teunissen [25] fit the measured
spectra of mechanically produced wind tunnel turbulence quite well
[41,42]. The spectral expressions for the streamwise and lateral velocity
components are given by
and
nS (n) __ u __ =
2 (]
u
where i = v and w.
4 (nLx/U) u
x nLi = 4(--) u
1+188.4(2nL~/u)2
(13)
(14)
For the comparison of these wind tunnel spectra with the normalized
x spectt:um functions, the turbulence integral scales, Li
, were determined
indepE!ndently using the previ.ously discussed direct method. The
-5/3 Von Karmon spectral equations show the correct n -dependence in
the inertial subrange. Integration of these expressions over the entire
48
frequency range leads to unity. At vanishing wavenumbers these
x I 2 spectral equations also predict the proper integral scale Li=Si(O)U 40i •
At low frequencies the logarithmic spectra vary as n+1 , which also fits
the wind tunnel spectra quite well.
However, caution must be taken with the automatic adaption of the
Von Karman spectral functions to atmospheric turbulence under near
neutral thermal stratification. The work of Kaima1 et a1 [20] clearly
points out that under near-neutral conditions the low-frequency content
in the spectra of the horizontal velocity components can vary appreciably.
In the convective boundary layer spectral data approach a definite shape
when near neutral conditions are approached from the stable regime.
Under these conditions significant low-frequency velocity fluctuations in
the u and v components are absent and the spectral shape of these components
is similar to the Von Karman model. On the other hand no unique spectral
shape for the u and v component exists for near neutral conditions
approached from the unstable regime. Under these conditions, the low-
frequency spectral content is much higher and the Von Karman model does
not represent the spectral data well.
In order to check the validity of the Von Karman spectral
functions, it will be necessary to check these relations against measured
spectra, where. the frequency is normalized with the local mean velocity
and the turbulence integral scale obtained independently via the direct
method. In Figures 42 and 43 logarithmic spectra of the u-component
measured at 5 different elevations for data record 19 (Table 1) are
compared with the Von Karman spectral function. The difference in those
two illustrations is that variances and integral scales used in Figure 42
49
were obtained in the middle-frequency range (0.00153-6.25 Hz) while
for Figure 43 the variances and integral scales were obtained in the
high-frequency range (0.02[.4·-100 Hz) filtering out the low-frequency
fluctuations. In the latter case the Von Karman expression fits the
measured spectra quite well in the -2/3 region but in the low-frequency
range the measured spectral values are considerably higher than the
predicted values. Consequently no distinct spectral peak at the
predicted value of the reduced frequency of nLx/U=O.146 is present. u
Similarly, the measured v spectra follow the Von Karman spectrum in the
~2/3 region only if variances and integral scales are used from which the
low-frequency components are filtered out. Figures 44 and 45 show
the w-spectra for run 1119 :In comparison with the Von Karman spectrum
function for variances and integral scales obtained in the middle and
high frequency range respectively. In the latter case the measured
spectra fit the theoretical Von Karman spectrum function much better
especially for the spectra from the higher elevations.
Based on these results the conclusion can be drawn that the theoreti-
cal Von Karman spectrum functions do not represent the spectra of atmo-
spheric turbulence in the low-frequency range or in the high-frequency
range~ when turbulence integral scales are used that are obtained via the
direct method when low-frequency fluctuations are included. Better fit
of the measured spectra in the -2/3 region is achieved when variances
and turbulence integral scales are used that are obtained from data
records from which the low-frequency fluctuations have been removed.
Inversely, if the theoretical Von Karman spectrum functions are used
to obtain streamwi.se turbulence integral scales,
50
x Li
, as is suggested in references 25 and 28, then these scales are
not equivalent to the integral scales obtained via the direct method.
The integral scales obtained via the Von Karman method must be inter-
preted as integral scales associated with velocity records from which
all low-frequency fluctuations with periods longer than approximately
40 seconds have been filtered out.
The expressions for the streamwise turbulence integral scales,
L~ for i=u,v and w, listed in references 25 and 29 are based on scales
obtained via the Von Karman method by either matching of the measured
spectra at the peak reduced frequency or by using a best overall fit
of the spectra. Consequently the predicted integral scales from these
sources must be interpreted as integral scales associated with the high-
frequency content of the turbulence components.
6.4.4 Comparison of Integral Scales Obtained Via the Direct and Spectrum Methods With Predicted Values from References 25, 26 and 29.
x . The turbulence integral scales, L , obtained from the cup-vane data
u
are obtained via the direct method in the frequency range from 0.00195 Hz
to 0.25 Hz. Integral scales obtained from the hot-film data and dis-
cussed in this section, are obtained by one of the following three
methods:
1. The direct method in the middle-frequency (MF) range (0.00153-6.25 Hz).
2. The Von Karman method, with the spectra and variances obtained from filtered data in the middle-frequency range (VK-MF).
3. The Von Karman method, with the spectra and variances obtained from filtered data in the high-frequency ra~ge (0.0244-100 Hz) (VK-HF).
51
For method 2 and 3 values of L~ were obtained by matching the measured
logarithmic spectra (nSi(n)/o~ versus nz/U) at nz/U=IO to the Von Karman
spectrum functions in the -2/3 range. Averaged and single-record stream-
x 'wise integral scales, Li' obtained from the cup-vane data or from the hot-
film data, the latter derived through one of the above three methods, are
shown for the two basic wind directions (south and northwest) in Figures
46 through 57 and are compared with the pred:lcted values of references
25, 26 and 29.
Figures 46, 47 and 48 show the variation of the averaged integral
scales, Li with i=u,v and w, obtained from the hot-film data (methods 2
and 3) versus height for southerly winds. The predicted values are based
on a roughness length, z =0.01 m in the IBL below 20 m and on a roughness o
length z =0.001 m above the IBL. o x The values of L from the cup-vane u
data match the ESDU [29] predictions and the scales obtained from the
hot-film data via the Von Karman method in the middle-frequency range
match the Teunissen [25] predictions. Values of LX and LX obtained v x
with the use of the Von Karman method in either frequency range fall
well below the predictions. Of course, as previously discussed, the
ratio of longitudinal scales and lateral scales is much higher for
southerly winds than for the other wind directions (Figs. 29, 30).
The effect of buoyancy in the inversion layer tends to suppress the
vertica.l motion and consequently the measured scales LX are much smaller w
than the predicted scales. It must be assumed that the predicted scales
are based on data records taken under conditions where buoyancy had
very H.ttle effect on the turbulence. The lateral scales, LX and LX v w
obtainE~d via the Von Karman method in the middle-frequency range are
about twice as long as the sca.les obtained from the same data in the
high-frequency range.
52
x Figures 49, 50 and 51 show the averaged integral scales, Li for
i=u, v and w, obtained from the hot-film data and analyzed according
to methods 1 and 2 for northwesterly wind directions. In Figure 49
comparison of LX is also made with the cup-vane data. The scales u
obtained via the direct method from the cup-vane data are generally
smaller than the scales from the hot-film data via the direct method
in the middle-frequency range. The scales obtained with the Von Karman
method in the middle-frequency range are systematically smaller than the
scales obtained with the direct method from either the cup-vane or
hot-film data in the same frequency range. The Teunissen [25] and
ESDU [29] predictions only fit the measured scales obtained via the
Von Karman method below a height of 20 m. Above this height the pre-
dicted values are $ystematica11y lower than the measured values. The
Counihan [26] predicted scales of LX fit the observed scales obtained u
via the direct method below 20 m, however above the height of 20 m the
Counihan prediction falls between the measured scales obtained via the
direct method and those obtained via the von Karman method.
The measured lateral scales LX and LX obtained via the direct method w v
in the middle-frequency range are three to four times as large as those
obtained via the Von Karman method in the same frequency range (Figs. 50,
51). The latter scales match the Teunissen [25] and ESDU [29] predictions
very well.
In Figures 52, 53 the integral scales, LX, obtained from a single u
early-evening record are compared with the predicted scales and with
the average scales from the daytime records all analyzed in the
middle-frequency range. It has been observed [13, 23] that just before
sunset the convective boundary layer dissolves abruptly and the low-
frequency velocity fluctuations normally associated with t~e convective
53
PBL suddenly disappear as can be clearly seen from the comparison of
the plotted records of the horizontal velocity components of record 16
(evening run) and record 19 (afternoon run) (Fig. 54). The results
indic.ate much smaller scales if the low-frequency fluctuations of the
horizontal components are absent. For record 16 the scales obtained via
the d.irect method are system.atically larger than those obtained via the
Von Karman method, although the difference is less significant for this
early evening run than for the daytime records, which contain low-
frequency fluctuations (Figs. 49-51).
x Figures 55 through 58 show the averaged integral scales, Land v
LX from the daytime records and those for record 16 obtained via the w'
Von Karman method in either the high-frequency range or the middle-
frequency range. The Von Karman method for obtaining integral scales
provides near-identical results independent of the frequency range if
largE! low-frequency fluctuations are absent (e. g. record 16) or if the
low-·frequency fluctuations are removed from the data records. If the
largE! low-frequency fluctuations are present (e. g. u and v components of
daytime records Fig. 54) and are not removed from the daytime records
the Von Karman method leads to much larger integral scales. If no
apprE!ciable low-frequency velocity fluctuations are present in the data
records, the integral scales obtained via either the direct or the Von
Karman method are about the same in magnitude. However, if low-frequency
components are present the magnitude of the scales depends ort the method
by which they were calculated and also depends on the frequency range in
which the data are analyzed.
54
When the integral scales, L~, obtained via the two-methods in
either one of the frequency ranges, for data records with or without
low-frequency content are compared with predicted values (Figs. 49, 52
and 53) one observes appreciable variation. In general, the Teunissen
[25] predicted values correspond to scales obtained from data records
for which the large low-frequency fluctuations are absent or are filtered
out, and obtained via either method. On the other hand the Counihan
[26] predicted values correspond more to the scales obtained from data
records with low-frequency fluctuations and obtained via the Von Karman
method in the middle-frequency range and to the scales obtained from
the cup-vane data. The magnitude of the scales obtained from the
same data records via the direct method are generally larger than
those predicted by Counihan [26]. The ESDU [29] predicted
scales fall between the two previously mentioned predictions. The scales
obtained from data records of north-west winds with low-frequency
fluctuations vary almost linearly with height, while the predicted scales
show more a tendency toward independency with height at higher elevations.
Similarly the lateral integral scales, LX (Figs. 50, 55 and 57) also v
show appreciable variation. The Teunissen [25] and ESDU [29]· predictions
correspond reasonably well to the scales obtained during daytime via
the Von Karman method in the middle-frequency range. The scales obtained
from the same' daytime records via the direct method are significantly
larger than the predictions. On the other hand if low-frequency compo-
nents are absent or filtered from the data records the measured values
for the lateral integral scales, LX, fall well below the predicted values. v
The integral scales, LX, (Figs. 51, 57 and 58) show a similar w
pattern, the ESDU [29] and Teunissen [25] predictions correspond best
55
to the scales obtained via the Von Karman method in the middle
frequency range. The scales obtained via the direct method in the
same frequency range are considerably larger than the predicted values.
If low-,frequency components are absent or r~noved from the data records
the mea.sured scales are generally lower than the ESDU [29] and Teunissen
[25] predictions.
In, general it can be concluded that the ESDU [29] and the Teunissen
[25] predicted LX and LX scales should be interpreted as scales obtained v w
via the Von Karman method for turbulence with low-frequency fluctuations.
X On the other hand the Teunissen [25] predicted Lu scales should be inter-
preted as scales obtained from data records from which the large low-
frequen,cyfluctuations are absent or are filtered out. In general the
horizontal scales L~ and L~ obtained from the daytime data records via
the Von Karman method in the middle-frequency range vary linearly with
height, while the predicted scales show a tendency of independnece with
height at the higher elevations.
6.5 Power Spectra
In this section the power spectra of the three velocity components
obtained from the hot-film data are discussed for the two basic wind
directions, south and northwest. Ample discussion of the spectra in the
previous section has indicated that the spectral shape in the low-frequency
range depends greatly on the absence or presence of large low-frequency
velocity fluctuations.
The logarithmic spe~tra obtained from south-wind records generally
show very little variation in shape and vary as f-2/ 3 in the high-frequency
~·l range and approximately as f in the low-frequency range. These two ranges
56
are separated by a distinct spectral peak. Kaimal (43) suggests that
under stable conditions in the absence of appreciable low-frequency .
fluctuations all spectra can be brought in coincidence and approximated
by the empirical relation
nS.(n) 0.164(f/f) __ ~1___ = _________ o~ __ ~
0. 2 1+0.164(f/f )5/3 1 0
with i=u, v and w,
where f=nz/U is the reduced frequency and f is the reduced frequency o
at the point of intersection of the extrapolation of the inertial
(15)
subrange of the spectra and the line ns i (n)/oi2=1. Kaimal's spectra and
variances are obtained from data in the frequency range 0.005~n~10 Hz.
The Von Karman spectrum functions (13, 14) can be modified into
similar expressions
and
with i=v and w.
nS (n) u
-2-= o
u
nS. (n) 1
0.156f/f o
= [
1+0.679(f/f )2 ] 0.12(f/f ) 0
o {1+0.255(f/fo)2}ll/6
In this set of equations the reduced frequency is defined as f=nL~/U.
However the parameter f/f for ,either the Kaimal or the Von Karman o
expressions represents the wavelength ratio A lA, where A is the o 0
(16)
(17)
wavelength associated with the reduced frequency, f. In both spectral o
expressions the parameter f/f is independent of the length scale which o
X is either the elevation or the turbulence integral scale Li
.
57
The Kaimal (15) and the Von Karman (16,17) spectrum functions are
n.early identical in the inert:lal subrange, but the Kaimal spectrum
predicts a slightly smaller spectral peak. In the low-frequency range
the u-spectra again are about identical, but the Von Karman v and w
spectra fall slightly below the Kaimal spectrum.
In Figures 59, 60 and 61 the normalized logarithmic u, v and w
2 spectra (nSi (n)/ai ) are plotted as a function of the modified reduced·-
frequency f/f. The spectra were taken from different data records for o
winds from southerly directions, which were classified according to the
local stability parameter z/L. The spectral data and the variances
were obtained from data analyzed in the high-frequency range 0.0244~n~100 Hz.
The velocity spectra obtained from stable-air records above the IBL do
n.ot differ from those obtained in the unstable air in the IBL and all
fit the Von Karman and the Kaimal spectral functions remarkably well.
The empirical spectrum functions (15,16,17) for estimation of the
velocity spectra in the case low-frequency fluctuations are absent can
be extremely useful if values of f can be predicted. Based on the o
experimental results it is obvious that f varies with height and with o
the presence or absence of appreciable low-frequency velocity fluctuations.
The results did not indicate any systematic variation of f with stability o
as suggested by Kaimal [43]. Averaged values of f =(nz/U) for each o 0
velocity component and obtained from normalized logarithmic spectra in
the high·-frequency range are shown as a function of height in Figure 62.
However, values for f obtained from the same data records but analyzed o
in the m:lddle-frequency range (O.00153.:sn~6.25 Hz) depend greatly on the
presence of low-frequency velocity fluctuations. If no appreciable low-
58
frequency fluctuations are present, the values f are independent of the o
frequency range the data are analyzed.
For all u-spectra investigated low-frequency fluctuations are
present between the low cut-off frequencies for the midd1e- and high-
frequency ranges, 0.00153 Hz and 0.0244 Hz respectively. These spectra
do not exhibit a distinct spectral peak but instead values of the
iogarithmic spectra are approximately the same for f<O.l. Values of
f obtained from the u-spectra analyzed in the middle-frequency range o
are generally smaller than those obtained from the same data records
but analyzed in the high-frequency range arid seem to converge to a
general value in the range 0.02<f <0.03, independent of height. o
For those records where internal gravity waves affect the spectra
above a frequency of n=0~00153 Hz, the frequency range for which
spectral values are increased varies with elevation. At the lowest
elevation (z=9.1 m) only the spectral values at the lowest frequencies
are affected and an appreciable range where the normalized logarithmic
+1 spectrum varies as f is still present (Fig. 63). However for spectra
+1 from higher elevations the f -range becomes gradually smaller as the
effect of the waves is felt at increasing frequencies until no appreciable
+1 frequency range with a f spectral distribution is present (Fig. 64).
For those cases values of f for the u-spectra seem to vary between the o
lower limit of f =0.02-0.03 and the values of f obtained in the high-o 0
frequency range as shown in Figure 62. Similar observations can be made
for the v and w spectra. The values of f obtained from spectra analyzed o
in the middle-frequency range are generally lower than those obtained
from the same spectra analyzed in the high-frequency range.
59
Spectra obtained from data records for westerly winds exhibit
significant low-frequency content for all elevations and the spectral
data fit the Kaimal or the Von Karman spectrum functions (15,16,17) in
the inertial subrange (-2/3 region) only (Figs. 65,66,67). The v-spectra
(Fig. 66) show a peculiar shape which is typical for spectra of the
lateral velocity component in the surface layer of a convective boundary
layer [19]. Kaimal's explanation for this shape is based on the fact
that in the inertial subrange the spectral values of the v-component
are 4/3 times larger than the u-spectra, a requirement for isotropy in this
range. (A similar situation exits for the w-spectra.) As the w-spectra
reach their peak and start to roll off with lower frequencies, the v'-spectra
instead continue to increase and start to follow the u-spectra. The
result of this is the peculiar shape of the v-spectra in the transition
between the -2/3 range and the low-frequency range where the u and v spectra
both are independent with elevation but instead vary with the height, zi'
of the convective boundary layer.
In the case the low-frequency fluctuations are absent as is the case
just before sunset, as the convective boundary layer.disintegrates rapidly,
the u, v and w spectra ,(Figs. 68,69,70) and specifically the v-spectra
(Fig. 69) have a completely different .character. The v-spectra show a
distinct spectral peak and a rapid roll-off at lower frequencies although
the spect,ra values fall above the Von Karman prediction in this range.
However in comparison with the spectra of the daytime run 19 [figs. 65,
66,67], the spectra of the evening run 16 show much lower spectra values
in the low-frequency range and the peculiar shape of the daytime v-spectra
as discussed above has disappeared and the v-spectra resemble the Von Karman
60
prediction.
Values of the reduced frequency, f , defined as the intersection o
of the extrapolation of the spectra in the inertial subrange and the line
n8 i (n)/oi2=1, were obtained from, spectra ~n~lyzed in both the high-
and middle-frequency range. For the daytime u and v-spectra for which
appreciable low-frequency velocity fluctuations are present, the values
of f obtained in the high-frequency range are systematically three o
times as large as the corresponding values obtained from the spectra
analyzed in the middle frequency range, while for the w-spectra the
ratio (fo)HF/(fo)MF is approximately 1.6, indicating that the 10w
frequency content is larger in the u and v spectra than in the w-spectra.
For run 16 this ratio for the u-spectra is 1.9 ~nd for the v ~nd w spectra
1.25. These results c1~ar1y indicate the effect of the 10w-frequen~y
fluctuations on the location of th~ inertial subrange when the spect~a
are presented in the logarithmic form with n8i
(n)/012 and f=nz/U as
coordinates. The distribution of fowith elevation for daytime spectr~
and evening spectra analyzed in the middle-frequency range are shown
in Figure 71. The results from the daytime spectra show that the values
of f for each velocity component are approximately independent with o
height and are (f ) ~O.Ol, (f ) ~O.02 and (f ) ~O.07. This observation ou ov·, ow
is in agreement with some of the results obtained fOr southerly winds
although the values are somewhat higher because of 1es§ low-frequency
content. If no low-frequency fluctuations are present in the velocity
components, the values of f generally increase with height (Figs. 62,71). o
The values of f can be used to obtain the wav~lengths corresponding o
to the spectral peaks associated with either the Kaimal or the Von Karman
spectral functions. Since for both empirical relations the logarithmic
61
spectral peak is approximately located at f/f ~3.8, then f ~3.8f and o m 0
for i=u,v,w. (18)
HI~re the wavelep,gth, Am=(U/l1)m.' corresponds to the peak of the Kaimal
and Von Karman spectrum functions and to the peak of the measured spectra
if no appreciable low-frequency fluctuations are present in the data
records. The peak wavelengths, (Am)i' are often used in micrometeorology
as measures of the energy containing eddies, or they can be slightly
modified to fit the original VOll Karman spectrum functions (13,14)
x from which values of Li can be predicted as proposed by Teunissen [25]
and
LX = 0.146 (A ) u m u
LXi = 0.106 (A ). for i=u,w m 1. (19)
However it must be realized that these predicted scales associated with
the empirical spectrum functions are equiv~lent to ,th~ turbulence
integral scale obtained from correlation functions only if no large
low-frequency velocity fluctuations are prel',lent. In tile case large
low-frequency velocity fluctuatio11s are present and not filtered from
the ~at~ 'f~cords, the scales qbtained from correlation functions are
gel1erally larger in magnitude (see section 6.4).
62
7. SUMMARY AND CONCLUSIONS
A rather detailed description has been given of a micrometeoro-
logical facility consisting of an instrumented 76 m (250 ft) tower
located within a 100 m distance from the shore at Wallops Island,
Virginia. The instrumentation system consists of cup-vane and
temperature instruments mainly used for profile measurements and a
hot-film system for turbulence measurements. The data acquisition
and handling system for the hot-film instruments is located in an
instrumentation trailer located at the base of the tower. The heart
of this system is a PDP 11/20. DEC minicomputer which controls the
digitization of the data (ZOO Hz sample rate) and the data transfer
onto digital tape. The digitized data have been analyzed on an IBM-
370 computer located on the VPI and SU campus.
Data have been acquired with the cup-vane system under moderately
strong wind conditions for all wind directions during a 4 l/2-year
period. From this data-base mean velocity and mean temperature pro-
files and associated parameters (roughness length, z , and power1aw o
exponent, a) have been derived as well as turbulence intensities,
x z z a Iu and a Iu, and turbulence integral scales, L ,L and L • u v u u v
Averages of the calculated flow parameters from all data records in
each 15-degree sector have been presented. In addition averaged mean
velocity ratios V/VZ50 ' turbulence intensities, au/U, av/U>and
turbulence integral scales, LX, have been obtained for 11 u
sectors each with near-uniform upstream terrain. The results provide
information about the microclimate at this site under moderately strong
wind conditions. This information is graphically presented in
63
Figures 72 through 75 from which average wind design data for this
coastal site can be established. In addition, data have been ac-
quired with the hot-film system for the southerly and northwesterly
prevailing wind directions. With this system turbulence parameters
such as turbulence intensities, Reynolds stresses, turbulence heat
x x x fluxes" integral scales (L ,L and L ) and power spectra of the u v w
three velocity components have been obtained.
Fc)r all observations made at this site under moderately strong
wind conditions, truly neutral thermal stratifications have never
been encountered throughout the observation height of 76.2 m for
any length of time. For westerly wind direction under sunny daytime
condit:lons the measured velocity and temperature profiles suggest
that the surface flow at the Wallops Island site is similar to the
surfac4~ flow observed during the Minnesota experiment [12]. The
observed PBL flow at Minnesota is an example of a typical convective
boundalt'y layer, a model of which is described in detail in Reference
13. In addition to mechanical and convective turbulence generated in
this atmospheric boundary layer, large-scale turbulence due to the
interaction of the mixed layer and the capping inversion (entrainment)
affects the mean and turbulent surface flow regardless of the wind
velocity. However, appreciable deviations from the convective
boundary-layer model may occur depending on atmospheric conditions,
time of day and wind direction.
It has been observed that just before sunset the daytime
boundary-layer flow is modified drastically as a result of the
64
disappearance of the large-scale turbulence and appearance of a
surface-based inversion. Similarly under conditions of low cloud
cover combined with precipitation, an inversion below the cloud
cover may develop, impeding the regular development of the day-
time convective boundary layer. Under these conditions the influ
ence of the large-scale turbulent motions on the flow below the in
version is reduced, resulting in an appreciable reduction in tur
bulence intensities and turbulence integral scales. Large negative
lateral velocity fluctuations or large wind direction fluctuations
towards the north have been observed for northwesterly wind direc
tions specifically between 300 0 and 315°. For winds in this sector
larger values of the lateral turbulence intensity and larger tur
bulence integral scales have been observed than for winds outside
this sector but with the same upstream terrain. In addition, tur
bulence intensities and turbulence integral scales vary during the
daytime as the convective boundary layer develops. The above
observations have been made under moderately strong wind conditions
with hourly mean-wind speeds between 10 mls and 20 mls at z;9.l m.
Based on all the observations made for westerly winds at Wallops
Island, there is no evidence that similar flow variations in the
surface layer would not exist under extreme and potentially damaging
wind conditions with velocities in excess of 20 m/s. Consequently
at this point in time it cannot automatically be assumed that for
extremely strong winds from westerly directions, the PBL flow at
the Wallops Island site is similar to the purely shear-generated,
65
neutrEl11y-stratified boundary-layer flow model, which is so often
advocated by wind engineers.
The mean and turbulent flow for southerly winds also differs appre-
ciab1y from that predicted by the neutral boundary-layer model. During
the summertime the warm air blowing over the cooler ocean water gives
rise to a surface-based inversion of variable height. Depending
on thE~ thermal stability, a low-level jet with a maximum velocity
occas:lona11y below the highest observation level has been observed.
Under extreme stable conditions the turbulence at the two highest
observation levels has been. observed to vanish completely and
generally internal gravity waves may co-exist with the turbulence.
Under these conditions the surface layer is very shallow, well below
the lowest observation 1eve1. Moreover, the flow near the surface
will also undergo a modification as soon as the ocean air crosses
the beach and experiences an increase in surface roughness and
surfalce temperature. TheSE! modifications of the surface flow
manifest themselves in the form of a developing internal boundary
layer (IBL) which at the tower location is between 15 m and 30 m in
height, depending on the change in surface temperature and the
overland development distance which varies with wind direction.
The conclusions of the boundary-layer experiment at Wallops
Island can be summarized as follows:
I. Westerly wind directions
1. The observed daytime flow below 76m at Wallops Island is described better by the convective boundary-layer model [13] than by the neutral boundary·-1ayer model.
2.
3.
4.
5.
6.
7.
8.
9.
10.
66
The height at which transition occurs from the logarithmic velocity profile to the mixed-layer velocity profile varies with wind velocity surface roughness and thermal stability.
The roughness length, z , obtained from velocity profiles below the transition arg in agreement with predicted values from PBL-flow review papers.
The Panofsky relation, a=z lIn Izl z2 , is only useful for predicting values of power~aw exponents for velocity profiles in the surface layer below the elevation where transition to the mixed-layer profile starts.
Measured turbulence intensities are generally in agreement with predicted values, except for the lateral turbulence intensities, a lu, in the northwesterly wind-direction sector 300 o <¢<3l5°. Yn this sector the turbulence intensities of the horizontal velocity components (u and v) are of the same magnitude.
Turbulence intensities of the horizontal components also vary with time of day and atmospheric conditions, or in general with the absence or presence of appreciable lowfrequency velocities fluctuating in the frequency range between 0.0015 Hz and 0.02 Hz.
The magnitude of the turbulence integral scales depends on the method (direct method or spectral method) by which they are calculated and also on the presence or absence of appreciable low-frequency velocity fluctuations.
If appreciable low-frequency content is present and is not filtered from the data records, the turbulence integral scales obtained via the direct method are larger than the predicted values.
The turbulence integral scales vary also with time of day, wind direction, surface roughness, and atmospheric conditions such as cloud cover combined with precipitation.
z z The measured vertical integral scales, Land L vary with direction of integration but are general~y in-agreement with predicted values.
67
11. The ratio LX/Lz varies generally between 3 and 4 and the u v
ratio LZ/Lz has an approximate value of 2. u v
12. The daytime turbulence spectra follow the Kaimal model [19] and deviate appreciably from the Von Karman model especially in the low-frequency range.
13. The Von Karman spectral model does not fit the measured spectra if appreciable low-frequeney velocity fluctuations are present and are included in the spectral analysis and in the calculation of the variance, ai' and integral scale (direct method).
II. Southerly wind Directions
1. For on-shore winds an IBL develops as the surface air passes the beach and experiences an increase in surface roughness and an increase in surface temperature especially in the summer during daytime.
2. For southerly wind directions warmer air flows over the cooler water creating a surface based inversion which is characterized by a very shallow surface layer, low-level maximum velocity (surface jet), low turbulence intensitie~ occasional vanishing of the turbulence under extreme stable conditions and co·-existence of turbulence and low-frequency internal gravity waves.
3. No simple boundary·-layer flow model is available to describe the on-shore flow at the Wallops Island site. Variations in observed velocity and temperature profiles, turbulence intensities and turbulence integral scales are extremely high and can occur within a very short time.
4. Measured velocity spectra (excluding the low-frequency gravity waves) are independent of thermal stability and seem to fit the modified Von Karman spectrum model (16,17) and the Kaimal stable spectrum model (15) extremely well.
As the above conclusions clearly indicate, there is no single and no
simple PBL-flow model available to describe the mean and turbulent
flow near the surface under moderately strong wind conditions at the
Wallops Island site. The presence or absence of appreciable low-
frequeI1Lcy velocity fluctuations causes the parameters describing
this flow to vary a great deal. The non--uniform surface conditions
68
and the presence of the low-frequency fluctuations, either in the
convective boundary layer or in the on-shore winds in the form of
internal gravity waves, cause the observed surface flow to be quite
different from the neutral boundary layer flow model. For the pre
vailing winds from either the south or westerly directions the
experimental results do not show any evidence for the PBL flow to
approach the neutral boundary-layer model as the wind speeq increases
from moderately strong to extreme.
It is assumed that engineers, already aware of uncertainty in
modeling the PBL flow, use safety factors in the design of wind
turbines to allow for differences in the actual wind environment
in comparison with the predictions from the neutral PBL model. The
variation of the turbulence intensities and turbulence integral scales,
measured under moderately strong wind conditions at the Wallops site,
is appreciable. Consequently a great deal of difference may exist
between actual measurements and the neutral PBL model. Experimental
evidence does not indicate that mean wind and turbulence parameters
will conform closer to the neutral PBL-model under higher wind-speed
and slightly unstable conditions. The observed differences appear at
the lower frequencies which are pertinent to the response characteristics
of the larger machines. Therefore, the design of large wind turbines
may need an increased safety factor with respect to turbulence at
frequencies below about 0.01 Hz.
69
REFERENCES
70
REFERENCES
1. H. H. Tie1eman and S. C. Tavou1aris, "A Method for the Measurement and the Statistical Analysis of Atmospheric Turbulence," NASA-CR-140586, 1974. Also Virginia Polytechnic Institute and State University, VPI-E-74-26, October 1974.
2. H. W. Tieleman and S. C. Tavoularis, "An Instrumentation System for the Measurement of Atmospheric Turbulence," Journal Industrial Aerodynamics, 2(1) (1977) 49-63.
3. H. W. Tieleman and D. B. Derrington, "An Experimental Study of the Atmospheric Boundary Layer Modified by a Change in Surface Roughness and Surface Temperature," Virginia Polytechnic Institute and State University, VPI-E-77-17, May 1977.
4. H. W. Tie1eman and W. W. L. Chen, "Statistical Analysis of LowLevel Atmospheric Turbulence," NASA-CR-137456, 1974. Also Virginia Polytechnic Institute and State University, VPI-E-74-3, January 1974.
5. H. W. Tie1eman, K. P. Fewell and H. L. Wood, "An Evaluation of the Three-Dimensional Split-Film Anemometer for Measurements of Atmospheric Turbulence," NASA-CR-62093, 1973. Also Virginia Polytechnic Institute and State University, VPI-E-73-9, March 1973.
6. H. Tennekes, "A Model for the Dynamics of the Inversion above a Convective Boundary Layer," Journal Atmospheric Sciences, 30(1973), 558-567.
7. J. W. Deardorff, "Numerical Investigation of Neutral and Unstable Planetary Boundary Layers," Journal Atmospheric Sciences, 29 (1972) 9l-ll5.
8. J. W. Deardorff, "Three-Dimensional Numerical Study of the Height and Mean Structure of a Heated Planetary Boundary- Layer," BoundaryLayer Meteorology, 7 (1974) 81-106.
9. J. W. Deardorff, "Three-Dimensional Numerical Study of Turbulence in an Entraining Mixed Layer," Boundary-Layer Meteorology, 7 (1974) 199-226.
10. J. C. Wyngaard and O. R. Cote', "The Evolution of a Convective Planetary Boundary Layer- A Higher-Order-Closure Model Study," Boundary-Layer Meteorology, 7 (1974) 289-308.
11. G. E. Willis and J. W. Deardorff, "A Laboratory Model of the Unstable Planetary Boundary Layer," Journal Atmospheric Sciences, 31 (1974) 1449-1452.
71
12. Y. Izumi and S. J. Caughey, "Minnesota 1973 Atmospheric Boundary Layer Experiment Data Report," AFCRL-TR-76-0038, Air Force Cambridge Research Labratories, Hanscom AFB, January 1976.
13. J. C. Kaimal et a1., "Turbulence Structure in the Convective BClundary Layer," Journal Atmospheric Sciences, 33(1976) 2152-2169.
14. S. J. Caughey and J. C. Wyngaard, "The Turbulence Kinetic Energy Budget in Convective Conditions," Quarterly Journal of the Royal ME!teorological Society, 105 (1979) 231-239.
15. S. J. Caughey, "Boundary-Layer Turbulence Spectra in Stable Conditions," Boundary-Layer Meteorology, 11 (1977) 3-14.
16. L. Mahrt et a1., "An Observational Study of the Structure of the Nocturnal Boundary Layer," Boundary-Layer Meteorology, 17 (1979)247-264.
17. S. J. Caughey, J. C. Wyngaard and J. C. Kaimal, "Turbulence in the Evolving Stable Boundary Layer," Journal Atmospheric Sc:iences, 36 (1979) 1041-1052.
18. S. Sethuraman, "The Observed Generation and Breaking of Atmospheric Internal Gravity Waves Over the Ocean," BoundaryLlllyer Meteorology, 12 (1977) 331-349.
19. J. C. Kaimal, "Horizontal Velocity Spectra in an Unstable Surface Llllyer," Journal Atmospheric Sciences, 35 (1978) 18-24.
20. J. C. Kaimal et a1., "Spectral Characteristics of Surface-Layer Turbulence," Quarterly Journal of the Royal Meteorological Society, 98 (1972) 563-589.
21. N. E. Busch, "The Surface Boundary Layer," Boundary-Layer Me:teorology, 4 (1973) 213-240.
22. S. J. Caughey and J. C. Kaimal, "Vertical Heat Flux in the Convective Boundary Layer," Quarterly Journal of the Royal Me.teorological Society, 103 (1977) 811-815.
23. H. W. Tieleman and S. E. Mullins, "Atmospheric Turbulence Below 75 m in the Convective Boundary Layer (Strong Wind Conditions)," Proceedings, Third Colloquium on Industrial Aerodynamics, Aachen, Depa~tment of Aeronautics, Fachhochschu1e Aachen, (1978) 1-23.
24. S. Sethuraman, "Structure of Turbulence over Water During High Wi.nds," Journal Applied Meteorology, 18 (1979) 324-328.
25. H. W. Teunissen, "Characteristics of the Mean Wind and Turbulence in. the Planetary Boundary Layer," UTIAS Review No. 32, UniveI'sity of Tqronto, October 1970.
72
26. J. Counihan, "Adiabatic Atmospheric Boundary Layers: A Review and Analysis of Data from the Period 1880-1972," Atmospheric Environment, 9 (1975) 871-905.
27. Engineering Sciences Data Unit, "Characteristics of Wind Speed in the Lower Layers of the Atmosphere near the Ground: Strong Winds (Neutral Atmosphere)," ESD Item No. 72026, 1972.
28. Engineering Sciences Data Unit, "Characteristics of Atmospheric Turbulence near the Ground, Part II: Single Point Data for Strong Winds (Neutral Atmosphere)," ESD Item No. 74031, 1974.
29. Engineering Sciences Data Unit, "Characteristics of Atmospheric Turbulence near the Ground, Part III: Variations in Space and Time for Strong Wind Conditions (Neutral Atmosphere)," ESD Item No. 75001, 1975.
30. D. M. Deaves and R. 1. Harris, "A Mathematical Model of the Structure of Strong Winds," ESRU Report No. 24, Environmental Sciences Research Unit, Cranfield, England, December 1976.
31. H. A. Panofsky, "Wind Structure in Strong Winds Below 150 m", Wind Engineering, 1 (1977) 91-103.
32. J. L. Lumley and H. A. Panofsky, "The Structure of Atmospheric Turbulence," John Wiley, New York, 1974.
33. H. A. Panofsky, "The Atmospheric Boundary Layer Below 150 Meters," Annual Review of Fluid Mechanics, Volume 6, Annual Reviews Inc., Palo Alto, California, 1974.
34. H. W. Tie1eman and S. E. Mullins, "The Structure of Moderately Strong Winds at a Mid-Atlantic Coastal Site (Below 75 m)," Proceedings Fifth International Conference on Wind Engineering, Colorado State University, (1979), 11-5, 1-15.
35. J. C. Kaima1, "NOAA Instrumentation at the Boulder Atmospheric Observatory," Proceedings Fourth Symposium on Meteorological Observations and Instrumentation, Denver, Colorado, (1978) 35-40.
36. A. P. van U1den, J. G. van der Vliet and J. Wieringa, "Temperature and Wind Observations at Heights from 2 m to 200 m at Cabauw in 1973," Royal Netherlands Meteorological Institute, De Bi1t, W. R. 76-7, 1976.
37. W. P. Elliott, "The Growth of the Atmospheric Internal Boundary Layer," Transaction American Geophysical Union, 39 (1958) 1048-1054.
38. H. L. Dryden et a1., "Measurements of Intensity and Scale of Wind-Tunnel TurbulenGe and Their Relation to the Critical Reynolds number of Spheres," NACA TR No. 581, 1937.
73
• 39. T. Von Karman, "Progress in the Statistical Theory of Turbulence,"
Turbulence, Classic Papers on Statistical Theory, S. K. Friedlander and 1. Topper editors, Interscience Publishers, Inc. New York, 1961.
40. J. O. Hinze, Turbulence, 2nd edition, McGraw-Hill, New York, 1975.
41. T. A. Reinhold, H. W. Tieleman and F. J. Maher, "Investigation of a Grid-Induced Turbulent Environment for Wind Tunnel Testing," Virginia Polytechnic Institute and State University, VPI-E-74-30, December 1974.
4·2. T. A. Reinhold, H. W. Tieleman and F. J. Maher, "Simulation of the Urban Neutral Boundary Layer for the Model Study of Wind Loads on Tall Buildings," VirgiI).ia Polytechnic Institute and State University, VPI-E-77-l2, March 1978.
4·3. J. C. Kaimal, "Structure Parameters in the Stable Surface Layer," Boundary-Layer Meteorology, 4 (1973) 289·-309.
74
FIGURES
,,~
.'"".,,",, ............. WAL' 'op S ISLAND ~ •• ", .• ',JV ... " '-
SC.~\L
? 'f' ';0 '1° ~o 100 It""
Figure 1. Map of Chesapeake Bay and Delmarva Peninsula
I
-...J U'1
'- --' -"
: \.:c
"
·10;. ... h"";~
'*-
\,Jo.:t .. r
..;.
\ 1.6S·
\ -' -. ...
.. .. Mo.. .... ~~
.... "'- ,,- .L.
At\Q.ntic. -Oc:.e.c..Y\ ~ ___ ./_~~t~~~-t.!~ ___ J~"!!.~_ ItWS
O \n.~o
J ,..
\,0''
i
"'T .. ~&. / ""
J.
o u
~o· , E."".-.t ,1S0
i,'S~
Figure 2. Wallops Island, Immediate Surroundings of the Meteorological Tower
'-I 0\
77
•. -Marsh, 335m "'-I 500m scattered brush
-Man:;h, 180m "'-I 335m scattered brush
". -Marsh, 180m scattered brush, bunker at 90m
-Mar~;h , 670m"'-l200m scattered brush
Parallel to Wallops Island
Ocean, Aerobee Tower 300 m upstream Prevailing
Ocean, 250m "'-I425m land South
over Wind$ -
Ocean, 500m "'-I 250m over land
In 'the sector 1400 < cp < 170 0 two rocket
assembly buildings 60 m ~ 150 m upstream
-Ocean,1l00m"'-l500m over land, some small buildings
Parcll\el to Wallops Island
Marsh, 500 m "'-I 850 m scottered small buildings and brush . -
I-
o <D rt>
0 0 rt>
OC/) vlLJ C\llLJ
a:::: (,!) lLJ 0 ..
06 00-- .....
(,) lLJ a:::: 0
0 OZ ~~
0 U)
o
til Q) H ;::l .j..J
C\I Q) ~
p .,-l OJ H H Q)
E-t
a OJ Q) H .j..J til p.. l=l
.t:: .j..J .,-l t3 til H 0 .j..J () Q)
en
I::l 0
.,-l .j..J () Q) H
.,-l t=l
"d I::l
.,-l t3:
. ('1'1
Q)
H ::l 00
',-l ~
78
Figu<e 4. The 76.2m (250 ft) Mie<ometeo<ologieal tower at Wallops Island
79
80
70
60 - RUN STARTING [3] TIME
0 7 14:12 50 - 0 8 15:37
6 9 17:05 E: 40 -... N
30 -
20 -
10 -
Ot~ ____ ~ __ ~~ ____ . __ ~~ __ ~ _____ ~ 295 296 297 298 299 300
PQTEN1"IAL TEMPERATURE, oK
Figure 5. Afternoon Temperature Profiles for Southerly Winds
80
60 I- \ I
~ 40~ ~
20 I-
0_3 -2
0 \ \ \0\ {
/).
\ \ \ 0 \
\/). \D
\ \
\\ \, \
\ \
~
-I o +1 +2 +3 +4 T- T oK o ,
Figure 6. Temperature Profiles for Westerly Winds. - - - Dry adiabatic lapse rate.
() U76 •2 = 16.7 mis, 12-21-'76, 12:22-13:39
o U76 •2 16.6 m/s. 1-26..,..'78, 18:52-19:43
o U76 . 2 = 20.9 mis, 8-9-'76, 17:07-18:07
~ U76 •2 = 15.9 mis, 1-27-'78, 01:09-02:09
C7 U76 . 2 13.8 mis, 1-27-t78, 05:42-06:58
00 0
+5
o It)
I ~1 I
(\J 0), > • ...... 0 o It)
> o -I-< a::: >-I--(.)
0 ....J UJ > % < UJ 1:
CD, •
0
tn, . 0
...,.
01 o 50
[!] MEAN (!) PSTD .t. HSTD + MAX. X t1lN .
100 ,
150 200 250 300
WINO DIRECTION, DEGREES Figure 7a. Variation of Mean-Velocity Ratio, V(15.2)/V(76.2) with Wind Direction
j 350
00 I-'
0 It) 0 t\I • >
" 0 0
> 0 -I--< a::
> CD .... • 0 -(.)
0 ..J IJJ > .....
• Z 0
< IJJ r
CD • o
tI)
•
~ ~
[!] MEAN (!) PSTD A MSTD + MAX. X MIN.
V A 1/
01 . I o 50 100 150 200 250 300
WIND DIRECTION, DEGREES
Figure 7b. Variation of Mean-Velocity Ratio, V(30.5)/V(76.2) with Wind Direction
00 N
350
N
''--1 ---:-1-------, o
lt) -t A MSTO /\ <'\I -1 + MAX. ~ -1 X MiN. -
o L{')
> o ...... f< 0::
>-r-...... u 0 ...J U.i > 00 . :z 0
< LLl l:
"-. 0
CD
j \/ /f
o~l __ ~--------~~r-~~----------~~--~--~~--------~--------~ o so 100 ISO 200 250 300 350
WIND DIRECTION, DEGREES
Figure 7c. Variation of Mean-Velocity Ratio, V(45.7)/V(76.2) with Wind Direction
CP w
o • i
o 11) It)
N 0
> :. ......... o o N > 0 o
• 0 -I-< 0:::
>-&I) 0)
I- 0 - 0
U 0 ....J UJ > Z < UJ E
II) CD · 0
0 co · 0 ,
0 50
[!) MEAN (!) PSTD A MSTD + MAX. X MIN •
100 150 200 250
WIND DIRECTION, DEGREES 300
Figure 7d. Variation of Mean-Velocity Ratio, V(61.0)/V(76.2) with Wind Direction
350
00 ~
E .. N
50
40.
30
10 I I
7 9 II 13 15 17 19 21 U m/sec
Figure 8. Typical Strong-Wind Profiles for Southerly Wind Directions
<Xl I..n
86
100~---R-'I--------~------------~ 90 zo,m
O -0.070 80°·031 70 0 -0.065 0.055
60
50
E ";30
20
15
~------------------~
IO~..-.-~~--~----~-----~~-------~ 12 13 14 15 16 17
U 1 m/sec
Figure 9. Typical Strong-Wind Pro{iles for Westerly Wind Directions Gradient Richardson Number (5) Evaluated at z=15.2m.
0.141 11.4
l 4 0 A- • I 0.12, ~1.2
&> 0.10 1.0
0
E O.ost A io.a E 0 0 A
":o.osl ~O.s .: 0 A
08 0 00
0 0 -...J
A 0.041-- @ 00 0 -t 0.4
en 0 0 0 0 A
0.02 I- 0 -t 0.2
cP 0 0' '0 230 250 270 290 310 330 350 10 30
WIND DIRECTION, DEGREES Figure 10. Variation of the Roughness Length z Versus Wind Direction
, 0' Note the larger roughness lengths in the sector 350 o <¢<+30°.
E .. 0
N
0.10
0.08 0
0.06 000
0 ())
0.04 & (jJ 0 0<:X:Q) 0
fo!- 0 0 0 0 0 OQ) 0.02 t- O
I-0 8~CDOO 9
0 ' , I I I I I " ""I • • , , , • I ,
0.001 0.01 0.1 1.0 -Ri
Figure 11. Variation 0:1; the Roughness Length, zoo with the Gradient Richardson Number(S) Evaluated at z=lS.2m
00 00
~ 0.26 ... i-Z W Z o 0.22 a. x w
~ --1 0.18 .-;;.. ~ .-
0:: W
~ 0.14 a..
O.IO'''''''''=: ",....-r .11 .'.1 ""
0.001 0.01 0.1 1.0 zo, m
Figure 12. Variation of Power-law Exponent with Roughness Length, Both Based on Velocity Measurements Below 45.7m. Panofsky's Relation (7) Based on zl=15.2m and z2=45.7m
c;:, \0
0.17 PANOFSKY [31J
- - - - COUNIHAN [26]
0.16 . ---ESDU [27]
I I
/ / ~ /
.. 0.15 I / l- I z / w o / z I 0
/ / a.. 0.14 x / w
/
0 1 ~ / ..J 0.13 / a:: / / w / / ~ / a.. 0.12
/ /
/ I ~ -
0.1 I 0.001 0.01 0.1
zo' m Figure 13. Variation of Power-law Exponent with Roughness Length,
Both Based on Velocity Measurements at all Levels. Panofsky's Relation (7) Based on zl=l5.2m and z2=76.2m
'" 0
eJ .. tZ W Z o a.. x w
~ --1
0:: W
~ a..
0.13
0.12
0.11
0.10
/
o 0
o o
~
o
0.09' .,,' X'"
0.001 0.01 0.1 zo, m
Figure 14. Variation of Power-law Exponent with Roughness Length, Both Based on Velocity Measurements Above 45.7m. Panofsky's relation (7) Based on zl=45.7m and z2=76.2m
__ Panofsky [31] --- Counihan [26] -- - ESDU [27]
\0 I-'
\I)
N
~ Z 15. 2M [!] MEAN
x I' (!) PSrD A MSrD + MAX.
=> 0-1 N &.~ X MIN.
........
>-I--........ (f) Z \I)
W I--Z ........
W <..J
~ oj \\ / II \!)
"" -1 => CD 0::: => I--
\I)
0 , , I I I . , 0 50 100 150 200 250 300 350
WIND DIRECTION, DEGREES
Figure lSa. Variation of Turbulence Intensity, cr /U, with Direction, z=lS.2m u
~rl------~----------------~
1" Z 300 5M I
1!1 MEAN I
~ (!) PSTD A MSTD
0 1 I + MAX. => Ni l X MIN • .......
>-~ ....... (f)
Z If)
W ~
Z .......
W (.)
~ 0 1 \\~ \\ f ¢ I It. -' => m a:: => ~
If)
0~1~~r-r-r-~~~~~-r~~~-'~r-r-~~~'-~~-r~~~~~r-r-~~~~~-r~ o 50
Figure I5b.
100 150 200 250 300
WIND DIRECTION, DEGREES
Variation of Turbulence Intensity, cr /U, with Direction, z=30.5m u
350
\.0 VJ
lI') (\J
~ Z 4S.7M (!] MEAN
;;-.e (!) PSTD A MSTD
::> ~.~ + MAX. X MIN.
t-4
>-~ t-4
en z lI')
W I-Z .~
W (.)
~ 0 1 ~\\~\ fJ) 1. / / I x
--1 ::> en 0::: ::> l-
ll')
0~1~~~-r~~~~~~~T-~~~~~~~~-T-r~~~~~~~~T-~~~ o 50
Figure l5c.
100 150 200 250 300
WIND DIRECTION, DEGREES
Variation of Turbulence Intensity, (J /U, with Direction, z=47.5m u
350
\0 .po
~Il----~==~------~----~
t ~ ~k~N
~ . (!) PSTD A MSTO
O~ + MAX. => (\J X MIN • .....
>-t-:-..... C/) z 11)
W t-:-:z ..... w (,)
~ol'/~\\' -l => en 0:::: => t-
11)
'" I r /I ~
o·~I~r-~~~-r~~--r-~'-~~~--~r-~~~~~~r-~~-r~~--r-~~-r~~--~r-~~· o 50
Figure l5d.
100 150 200 250 300
WIND DIRECTION, DEGREES
Variation of Turbulence Intensity, a /u, with Direction, z=61.Om u
350
\0 VI
I/) N'
j Z 76.2M [!] MEA~
~ (!) PSTD A MSTD
• 0-1 + MAX. => N X MIN. .....
• >-t-..... CJ) Z I/)
W t-Z ..... W (,)
Z 0 1 W
/ \ \\ l'f /I V' \tl ...J ::> m cr => t-
O;I~~~~~~~-r-r~~-r-r~~~~~~'-'-'-~~T-T-~T-T-~~~~~~ o 50
Figure l5e.
100 150 200 250 300
WIND DIRECTION~ DEGREES
Variation of Turbulence Intensity, cr /U, with Direction, z=76.2m u
350
\0 (j\
~
• > -
• >-I-.... en z l1J t--:z: -w (.)
Z l1J .....J :::> co 0:: :::> I-
0, N
0, -
J I ~~=I ·/-tl 1 (!) PSTD
A HSTO + MAX • X MIN.
o~I~~~~~~~~~~~~~~~~~~r-r-r-r-r-r-r-~?-?-?-~~~~~~~ a 50
Figure l6a.
100 150 200 250 300
WIND DIRECTION. DEGREES Variation Turbulence Intensity, a /U, with Direction z=lS.2m
V'
350
\.0 '-I
0 N I I Z 30.SM
I!JMEAN ;N! J (!) PSTD
A MSTD + MAX.
> I X MIN. .......
• tI) >-t-....... C/)
z W t-Z .......
W 0
U z 1 \\ \ \ \ / i III w -1 ::::> CD a::: ::::> t-
tl)
O~'~~~r-r-r-T-~~~~~~~~~~~~~~~~~~~~~r-r-r-T-~~~~~~ o 50
Figure 16B.
100 150 200 250 300
WIND DIRECTION. DEGREES Variation Turbulence Intensity, (J lu, with Direction z=30.5m
v
350
\.0 00
l Z 45.7" I .. ~/ll 01 [!] MEAN ~ I N (!) PSTD i ~ A MSTD I
I + MA X • I / ,!II. "I I > t-4
>- tf) I-t-4
(f)
Z W IZ t-4
W 0 U Z W --.l => CD c:: => I-.-
tf)
L XMIN. /\ '¥ pi I i
I I ! ,
0~1--~~~~-r~~--~T-'-~~-'~~r-~~-r~~r-r-~~-r~~--~T-'-~~~--~T-'-4 o 50
Figure l6c.
100 150 200 250 300
WIND DIRECTION, DEGREES
Variation Turbulence Intensity~ cr /U, with Directionz=47.5m v
350
\0 \0
0 N
J I Z 61M (!] MEAN
~ I (!) PSTD A MSTD + MAX.
> 1 I X MI N. .......
II) ->-..-....... (f)
Z W ~
Z .......0 -W U Z
~.{ \\~ / J II 'Y. XI w .....J :::)
CD 0::: :::)
..- II)
O~I~~~~r-r-r-r-r-~~~~~~~~~-r-r-r-T~~~~~~~~r-r-r-r-~~~ o 50
Figure 16d.
100 150 200
WIND DIRECTION. 250
DEGREES 300
Variation Turbulence Intensity, cr lu, with Direction z=61.Om v
350
I--' 0 0
~
> .......
>-I-....... (f)
Z W I-Z
a N-I--------------------------------------------------------~
II) -
1
- Z 76-2M [!] MEAN (!) PSTD A MSTD + MAX. X MIN.
....... a -W U Z w
1'- 'bt~ \ / / // xl -1 :::> ,..,... 1...1..1
0::: :::> I- II)
O~I~~_r~--r_~~~~~_r~--r_~~--r_~~~~~~~--r_~~~~~_r_,--r_~~--r_~~~~~~
o 50 100 i50 200 250
WIND DIRECTION. DEGREES
Figure l6e. Variation Turbulence Intensity, 0 lu, with Direction z=76.2m v
300 350
...... 0 ......
80 o 2400~ cp~ 2550 ,0 l:l 0 2400~ cp~ 2550
\ \ , \ \ l:l 3000~ cp ~ 3150 l:l300° ~ cp ~ 3150
\ \ \ '0 l:l 60~ \ \ \ \ \ \ \ \ \ \0 0 l:l
~ 40~ \ \ \
\ \ \ \ .....
'0 '? 0
l:l N
\ \ \ \
20~ \ \ , , ,
\ , , 0 l:l " " ,
..... ..... " ..... ..... .....
I 0
13.0 15.0 17.0 19.0 21.0 90 11.0 13.0 15.0 17.0 19.0 o"u/u, 0/0 O"V/U,%
Figure 17. Variation of Average Turbulence Intensities a Iu and a Iu with Height for Two Wind-Direction Sectors. u v -- ESDU [28] ------Teunissen [25]
103
Direction Magnitude
N'-W Wind Records from an Aerovane a1' ! = 10 m
Figure 18. Typical Strong-Wind Aerovane Record of Direction and Speed for Northwest Winds, CP~320o, U10~13 mls
104
0.5~----------------------------------~
~ ~ 0.4 0- • >- ~lII t--~ 0.3 lJ.J Cl • >-50 .2 -CD « CD 0 0.1 a: a..
o 1--____ ....... I!IIE..-4----...L.---~ • __ ~I111*0-----L
-6.0 -4.0 -2.0 0 +2.0 +4.0 <p / 0-;
PROBABILITY DENSITY FUNCTION WIND DIRECTION. 0-; = 8°
Figure 19. Probability Density Function of Wind-Direction Fluctuations, z=76.2m, cr<b=8°, ¢=322° ------ Gaussian Distribution Function
+6.0
ao~~~~==~==============----~----------------~
I \ b. g :~~:OON I \ ~ NIGHT \ no ~ - ESDU [28J \ 0 COUNIHAN [26J
\ --- TEUNiSSEN [25J
60
~40 '00
\ \ N
" <>
20
" 00 " "-""-,
............ ____ 0
--------f:::t.
O~I----~--~----~--~----~--~--~~--~--~ 12 13 14 15 16 17 18 19 20 21
o-u /U, 0/0
Figure 20. The Variation of the Vertical Distribution of Turbulence Intensity a lv, with Time of Day for Westerly Winds u
I-' o Vt
106
80~-------------------------A
60 A ~
\ \ \ \
E sou [28J
TEUNISSEN [25J
HOT FILM
E .. 40 \ \
N
20
\ \
\
6 \ \ , , , ,
8, .....
"-6 .... ~
0---............... -----'-----'------' 6 9 10
Figure 21. Vertical Distribution of Turbulence Intensity, a lu, for Westerly Winds w
107
8e) 61D
-ESDU [28J 7C) o AVERAGE OF 9 RUNS
(HOT FILM) 6 RUN 7 (HOT FILM)
6CI lD o AVERAGE OF 29 RUNS
(CUP-VANE)
50 600
E .. 40 N
30 6 00
1 STABLE
20 IUNSTABLE 000
10 18 L
OL-______ ~I~--.----~I----------~--------~ o 5 10 15 20
o-u / U, 0/0
Figure 22. Vertical Distribution of the Turbulence Intensity, a lv, for Southerly Winds u
108
80 - ESD U [28J OJ
o AVERAGE OF 9 RUNS 70 (HOT FILM)
~ RUN 7 (HOT FILM)
60 <1l o AVERAGE OF 29 RUN (CUP - VANE)
50 6 aJ
E.-40 N
30 6 0 0 STABLE
20 UNSTABLE
IBl 10 0
o--------------~--------~------------~ o 2 4 6 8 10 12 14 (jv I U, 0/0
Figure 23. Vertical Distribution of the Turbulence Intensity, a ju, for Southerly Winds
v
16
109
50 60
E .. 40 N
30
20
10
o 6 0
o
STABLE
UNSTABLE IBl
o
O~--~I~--~--_~,----~I--_--~--~----~--~ o 2 4 6 8 10 12 14 16
Figure 24.
crw / U , %
Vertical Distribution of the Turbulence Intensity, a /u, w for Southerly Winds
1.05 • i
... {-Q- CUP-VANE
o HOT FILM
0;95
b:::S
" ~0.85 0.70
~ ~ 0
0.75 A
A A aAA
I 0.60 ~q A .......
I f=q
-HOT FILM A • 0.65
0 60 120 180 240 300 368.50
WIND DIRECTION,· DEGREES
Figure 25. Variation of Ratios of Average Turbulence Intensity with Wind Direction
o ~~I------------~i------I------------------------------~
I
Z 1 S. 2M I" ~ 1!1 MEAN
(!) PSTO · 81 A MSTO x II) + MAX. => X MIN • ...J
• LtJ ...J 0
0 < • (.) en
• -< Q:: 0 (!) 0 LU ", I-Z
I / \ /J\\ /~ \ -• 0 m 0
Q:: N
=> I-
0 0-
o'~I~~~~ __ ~~~ __ ~~~ __ ~~~~ __ ~~~~ ____ ~~~ ____ ~~~~~ o
Figure 26a.
50 100 150 200 250 300
WIND DIRECTION, DEGREES Variation of Turbulence Integral Scale, LX, with Wind Direction, z=lS.2m
u
~ ~ ~
1:
• x => -l
• lJJ -l < (,) en -l < a::: (!)
lJJ .-z -
• m a::: => .-
0 0 CD
I Z JO.SM
~ 1!1 MEAN (!) PSTD A MSTD + MAX. X MIN.
0 0 ~
0 0 ",
1 -. , - , I I 1 , \1 , , \ I
01 ""- I / "'\ I I '\. ... I / \ r If. tI\.. U~I 0 N
o~t~-T-T~~~~~~~~~~~~~ __ ~~~~~~~~T-T-T-~~ ____ r-~~
o
Figure 26b.
50 100 150 200 250 300
WIND DIRECTION. DEGREES
Variation of Turbulence Integral Scale, LX, with Wind Direction, z=30.5m u
350
......
...... N
o g'I----------------------------------------~--------~-------------
~
• x :::> --l .....
• UJ --l < o (/)
--l < ac <.!) UJ tZ -m ac :::> t-
i
o 50
Z 45.711 [!) MEAN (!) PSTD A MSTD + MAX. X MIN.
150 200 250
WIND DIRECTION, DEGREES JOO
Figure 26c. Variation of Turbulence Integral Scale, LX, with Wind Direction, z=4S.7m u
350
~ ~ w
0 0 Q)
I Z 61M
1:8-1 [!J MEAN (!)PSTD
~ nl • MSTD + MAX. X MIN.
-J ""-
• IJJ -J 0 < 0 OCD en -J < a:: (!)
IJJ .-~ , I
• 0 /r\\ ! \1/ "'VI" \VJ j \~ J" CD a:: .::> .-
O~I~~~~~~~~~~~~~~-r-r~~~~'-'-'-~'-~~~~~~~~~~ o 50
Figure 26d.
100 150 200 250 300
WIND DIRECTION. DEGREES Variation of Turbulence Integral Scale, LX, with Wind Direction, z=6l.Om
u
350
I-' I-' ~
o ~~I------~~-----------------------
1: ~- ~ ~A~ II CD (!) PSTD
• A MSTD x g.J + MAX .. J ::> 0 X MIN. -l -
• LLI -l < (.)
en -l < 0::: (!)
LLI ..... Z -
• 0 m 0 a::: ", ::> .....
o o N
o o
O~I~~~~~~~~~~~~~~~~~~~~~~~~~~~r-r-~T-T-~~~~ o 50
Figure 26e.
100 150 200 250 300
WIND DIRECTION. DEGREES
Variation of Turbulence Integral Scale, LX, with Wind Direction, z=76.2m u
350
I-' I-' I..n
~,
I
1: oj
• I +-N :::> ...J
-UJ ...J < 0 en ...J < 0:: (!) lLJ ~
I , )\\, Z ..... •
m 0:: :;:) ~
o 50
Figure 27a.
Z 15.2K I I!J MEAN
(!) PSTD I .& MSTD + MAX. X MIN.
II / \/\ 1\\ I I' / \ I/o
100 150 200 250 300
WIND DIRECTION. DEGREES Variation of Turbulence Integral Scale, LZt, with Wind Direction z=lS.7m
u
\\ I f-' f-'
'"
350
I ; I --- _. -.~-
1: [!J MEAN (!) PSTD
.f-
1- I" Z 30.SH i
.. 0 I I A r;sTO "'1 + MAX • - X MIN.
N ::> -J
• l1J -.oJ < 0 CJ)
...J < 0::: 0 UJ I-Z -•
m 0::: => I-
0 ~,
~ .. 4" --.
011~~~~~~~~'-~~T-~~~~~~~~~-T~~~~~~~'-~T-~~ o
Figure 27b.
50 iOO i50 200 250 300
WINO DIRECTION. DEGREES Variation of Turbulence Integral Scale, LZt, with Wind Direction z=30.5m
u
I-' I-' -...J
1:
• 0 II)
+-- -
Z 45.7H [!] MEAN (!) PSTO A HST"O + MAX. X MIN.
N ::::> ...J
• UJ ...J < (.)
en ...J < ~ (!) lLJ r-:z -
• CD ~ ::::> .....
0 0
~ ~
01 • o 50 100 150 200 250 300
WIND DIRECTION. DEGREES
Figure 27c. Variation of Turbulence Integral Scale, LZt, with Wind Direction z=4S.7m u
350
......
...... (Xl
o
:~1------------1-7-_-45-~-7"-.1-------------------------~----~I
:E: =1 I m ~.~~ I j" "\ I" • g.J • "STD - " + MAX. + . . .... -N => 0. ...1 Ot
• llJ O . ...1 CD
<" (.) OOe· -J <0 cr CD' C!) llJ toZ .... CD cr :> .-
XI11N.
~~
o 50 100 150 200 250 300
WIND DIRECTION. DEGREES Figure 27d. Variation of Turbulence Integral Scale, L:t, with Wind Direction z=4S.7m
350
I-' I-' \0
i' I
Z 45.7"
~ ~1 t!l MEAN (!) PSTD A MSTD
" + MAX. X tUN.
NO ::::> .... ~
• UJ ~
< (.) en ~
< ct: (!) UJ l-
I l!I \. / , - \ 111'\. / "-l )t-- .... -.,cl Z ~ - 0,
• II? til 0::: => I-
o I j I o 50 100 '50 200 250 300
WIND DIRECTION, DEGREES Figure 27e. Variation of Turbulence Integral Scale, LZi, with Wind Direction z=4S.7m
u
350
~ N 0
0 N -0 -.
E
• ......... N ~ ....I
• W ....I < (.) (f)
....I < a=: (!) LU to-Z -
• CO a=: :::> t-
O N
O.
0 I
0 50
Figure 27f.
1 z 61.0~-1
I· [!] MEAN I (!) PSTD .. MSTD + MAX • I X MIN. I
100 150 200 250
WIND DIRECTION, DEGREES 300
Variation of Turbulence Integral Scale, LZ~, with Wind Direction z=6l.0m u
350
I-' N I-'
0 N -0 -r
• --+
N ::> -l
• L&J ...J < (.) CJ)
...J < 0:: (!) LU ~ % -•
al 0:: => ~
0 ('Ii
O. -o . • 0 50
Figure 27g.
'Z 78.2" [!J ttEAN (!) PSTD ... I1STD + MAX. X MIN.
100 150 200 250
WIND DIRECTION, DEGREES 300
Variation of Turbulence Integral Scale, LZ"" with Wind Direction z=76.2m
u
350
..... N N
1: :~l_ I ~ ~i.~ I I~\ I m ~ PSTO ~_
~ A MSTD / ~
I + MAX. I I V 4--
N > ...J
~ g ...J. < (,)
(J) 0 It)
.....J < 0:: (!) 0 lJJ ~ r-z - 0 . .." en 0:: :::::> r-
I \
o t 1-
o
Figure 28a.
X MIN. t
• / A\/I :J A
50 100 150 200 250 300
WIND DIRECTION, DEGREES Variation of Turbulence Integral Scale, L
Z +, with Wind Direction z=lS.2m v
\1
350
f-' N W
L
.. .--N > -'
.. UJ -' < 0 en -' < a:: (!)
1JJ t-:z -
• Ol a:: ::::> I-
0 N -0
0 Il)
0 .. o. .., o. N
O.
0,
0 50
Figure 28b.
Z 30.5" [!) MEAN (!) PSTD A HSTD + MAX • X HIN.
100 ISO 200 250
WIND DIRECTION, DEGREES 300
Variation of Turbulence Integral Scale, LZt, with Wind Direction z=30.5m v
I-' N ~
:1 i Z45.7H I ~ ~ I 1:. ~ [!] MEAN I \ I \ - ~ PSTD
• 0 A "STD +- ~-I ~ Ht~: . . N > O. --l ~
• lJJ ~. --l < (.) en -' <: 0:: (!) UJ
~ 11)1 - 0 ••• en 0:: ::> I-
.\ f/\\ , \ IT )It v- I If ""
o~I~ ______ ~~~ __ ~ __ ~ ______ ~~ __ ~ ____________ ~ __ ~ _____ , ____ ~ 150 200 250 o
Figure 28c.
50 100 300
WIND DIRECTION. DEGREES
Variation of Turbulence Integral Scale, LZt, with Wind Direction z=45.7m v
350
~ N V1
g. I
Z 45.7M
~ gi I!J MEAN (!) PSTO A HSTD
;; ~I + MAX. X MIN.
> ..J
.' IJ.J .J < (.)
en 0 11)'
...J < cr (!)
LLl .....
~gt \\f T\ \ I \ \VJN /\/\ I-' I'-) (j\
CD a:: ::)
..... 2·
0, -°1 ' • o so 100 150 200 250 300 350
WIND DIRECTIONw DEGREES Figure 28d. Variation of Turbulence Integral Scale, L:t, with Wind Direction z=45.7m
g~I------------------------------------------------------------~ - I
~ l 45.7" I
l: 0 [!] MEAN ~ m PSTO
• . 6 KSTD --+ N > .-oJ
• lLJ .-oJ < 0 en -' < ~ (!) 0 lLJ .,..' .-z -
• CD 0:= :::> .-
+ MAX. X tUN ..
I " I
" 1/\\
50 100
\ / 'Y
150 200 250
WIND DIRECTION, DEGREES
JA • V ~I
Figure 28e. Variation of Turbulence Integral Scale, LZ~, with Wind Direction z=4S.7m v
I-' ~ -....!
0 No -
l:~ ~ 61.0" (!) MEAN (!) PSTD A MSTD
-.. ~I N
I + "AX. X HIN. > ..J
• IJJ o . ..J CD
< (.) en ...J < ex: C)
UJ
~ ~1 .\ If \\ / "4/ j, V'\ I - 0, • •
m a:: ':) 0, ..... .."
O~I~-T~-r~~~~~~~~~~~~~~-T~~-r~~~~~~~~~~r-~ o
Figure 28f.
50 100 150 200 250 300
WIND DIRECTION. DEGREES Variation of Turbulence Integral Scaie, LZ~, with Wind Direction z=6l.Om
v
350
I-' N 00
r-----~ _ i
1: l!J MEAN 1 Z 1&.211 I" (!) PSTO
• g I A "STD --+ ~ + MAX. N > ....J
• tJJ -l '< () en 0
....J < a::: 0 UJ I-% -
• m a::: ~, :) J-
X tlIN.
~ , ~
01 r b 50 tOO J50 200 250
//
WINO DIRECTION, DEGREES
AV
300
Figure 28g. Variation of Turbulence Integral Scale, LZ+, with Wind Direction z=76.2m v
-I I-' N \0
350
N :::> ..J
" X :::> ..J
• .. -r.-------------------------------------------------------------------------------------------------------------,
..
.. , ..
[!) Z.'5.2M t (!) Z=30.5M t A Z=45.7M l' + Z-45.7M ~
N I i o
Figure 29.
50 100 150 200 250 300 350
WIND DIRECTION, DEGREES Variation the Integral-Scale Ratio LX/L~ for the Three Lowest Observation Levels with Wind Direction. Arrows Indicatg the Direction of integration.
..... w o
N :::> -J ........ x :::> ...J
, I!I 2-45 -7M t --l ~
G Z-45.7M + ... Z-SO.1M 4 ~ + 2-76- 2M .a. .
O~I-T-r~~~~-T-r~~~~-T~~~~~~-T~~~~~-T-r~~~~ o
Figure 30.
so 100 1 SO 200 250 300 350
WIND DIRECTiON, DEGREES Variation of the Integral-Scale Ratio LX/Lz for the Three Highest Observation
u u Levels with Wind Direction~ Arrows Indicate the Direction of Integration.
I-' W I-'
80. I
/ I 0 6. I I
I
/ , 60~
I 0 6. I , I
/ I I I I 0 6.
~ 40t I / I I I I
I /A 0 I-'
I w .....,
I 0 2.40Q < <p < 2550
20~ I I
I ~ 6. 300 0 < <p < 315 0
I ESDU [29J I /
I / --- COUNIHAN [26] I
0- _ ----- TEUNISSEN [25]
0 50 100 150 200 250 300 350 400 450 x
LU' m
Figure 31. Variation of Average Integral Scales LX (Cup-Vane) with Height for Two Westerly Wind-Direction Sectors u
I
60 ? ~ I
E 40 .-~ '? ~ I
N
M I
/ /
6 ! / o 2400 < cp < 255 0
./ ~ 3000 < cp < 315 0
/ - ESDU [29J
20
- -- COUNIHAN [26] ~ I I O~I------~----~~----~----o
Figure ~2.
25 50 75 100 125 150 175 z
Lu,m Variation of Average Integral Scales, LZ
, (Cup-Vane) with Height for Two Westerly Wind-Direction Sectors. A¥rows on Symbols Indicate the Direction of Integration
I-' w w
80 o 2400 < </> < 255 0 I 6. 3000 < </>< 3150 9 / ~
-ESDU [29J
60~ 9 tf
! E ... 40 1 9 R 6r + N ,
6 !
20~ /6 !
0' , o 10
Figure 33.
20 30 40 50 z
Lv,m 60 70 80
Variation of Average Integral Scales, LZ, (Cup-Vane) with Height for Two Westerly Wind-Drecition Sectors~ Arrows on Symbols Indicate the Direction of Integration.
..... w .po
8°1 60
I I I I I I I I I
I I
I o I I
o I ,
o I E~ 40 ! N
20
I I
I I
I I
I /
/ I
I
°0 50
Figure 34.
/ I
/ ,/
100
o STABLE
UNSTABLE IBl
150 200 X
LU,m
I -- E.SDU [29] ---COUNIHAN [26] ---- TEUNISSEN [25]
250 300 350
Variation of Average Integral Scales, LX, (Cup-Vane) with Height for Wind-Direction Sector 1800 -195 0 (South Winds) z<2o.m+z gO.Olm and z>2Om+z =O.OOlm
·00
I-' W VI
80. · ? / I
J
60 « / I , E 40 .. ~? / I
I N
6 / 20
/ , /
, STABLE
/ ---.' I
I
J
UNSTABLE I BL "-1 --E-S-O-U-[-2-9]--t
-·-COUNIHAN [26] o ' · o 20
Figure 35.
40 60 80 100 120 14·0 160 z
LU,m Variation of Average Integral Scales, LZ
, (Cup-Vane) with Height for Wind-Direction Sector 180°-195° (So3th Winds). z<20m+z =O.Olm and z>20m+z =O.OOlm.
o 0
Arrows on Symbols Indicate the Direction of Integration
..... W 0\
60 ~
9 &} E 40 ... . N
6 20 6
00 10 20 30
STABLE
UNSTABLE IBL
-ESDU [29J
40 z
LV,m
50 60 70 80
Figure 36. Variation of Average Integral Scales, L~, (Cup-Vane) with Height for WindDirection Sector 180°-195° (South Winds) z<2Om+z =O.Olm and z>2Om+20m=O.OOlm Arrows on Symbols Indicate the Direction of Inte~ration
..... V,) ,
.......
138
80~--------------~~----~
60
E -- 40
N
20
"
Figure 37.
o MORNING IJ AFTERNOON A NIGHT
200 x 300 Lu,m
400
x Profiles of Average Integral Scales,Lu' for Westerly Winds for Different Times of the Day.
350r~-------------------o~------~·----·-
I \ I ESDU [29] I . '.. 0 ~--- COUNIHAN [26J
300 ~ \ IL~=43m TEUNISSEN [25J I ,
250 t- '" o
, .'" o
o o ~ 200 I-
o " 0
)(:J .-I
150 t-
100 I-
500.001
Figure 38.
o
I
0.01
" 0 0 " 0
00 b.(i,,~ o <6 ~
<9 0 '~ 0 0
o·
o o ......
~O o o 0 - -0.
..1
0.1 zo,m
1.0
Variation of the Integral Scale, LX, at z=15.2m (50 ft.) with Roughness Length, u
zo' for all Data Records with a Westerly Wind Direction and Ri15
>-0.1
.... Vol \0
6oolr--------------------------~--------~--~ . 0 I ESOU [29] I I . - -- COUNIHAN [26]
500 I- 0 I ---- TEUNISSEN [25] I 00 ·
400 o o
E cP 0
)( ; 300 --1
'~ 0 0 0 0 ... ~ 0 00
----_~ 0
200 o 0 _______ _
00 0 o--ct o ~-----------------~ r
100
o ' , 0.001 O.Ot 0.1 [0
Figure 39. ~,m
Variation of the Integral Scale, LX, at z=45.7m (150 ft.) with Roughness Length, zo' for all Data Records with a We~terly Wind Direction and Ri
15>-O.1
I-' ~ o
E .-.....
9°1· 00 ~ I-ESDU [29J I - tJ 80l 0
70 0
<e>
0 0
0
c9 0
o o 0 o
'~60 o 0 0 <0
.50·~
40
3°0.001
Figure 40.
000
0 o 0
o
o o
o 0.01 0.1 1.0
zo,m Variation of the Integral Scale, LZt, at z=45.7m (150 ft.) with Roughness Length, zo' for all Data Records with Weste¥Iy Wind Direction and Ri15>-O.1
I-' .p-I-'
70 I » \ ,
-ESDU [29]
60
0 I
50 E ~ 0 I ...
0 ~40~ 0 0 no 0
o~ ............... J
0 30 t- O 00 @>o o 0
0 ... · 0
0 20t
0 0 0
10' , 0.001 0.01 0.1 1.0
Figure 41. zo, m
Variation of Integral Scale~ LZt~ at z=4S.7m (150 ft.) with Roughness Length~ zo~ for all Data Records with v Westerly Wind Direction and Ri 15>- 0.1
I-' .j::- . N
0.00 I i
-0.50 (!)
~ -1.00 I (!)A J (!JA: b::l , -C -:J -1.50 en c
'-r-I
/ \.~ (.!) 0 --l -2.00
-2.50
VON KARMAN SPECTRUM (13)
-3.0~3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 4.00
LOG (nL~ /U) x Figure 42. Logarithmic u-Spectra for Run 19, cr and L Obtained in the Middle . . . . x u u
Frequency Range, L Obtained Via the Direct Method u
I-' ~ w
0.00 .--. ---------------------,
-0.50
XX A'" rn?< X
l!l 1<
A 6< li!l ...
Cl
~ -1.00 I- Cl
t? " -C
-:3 -1.50 en c .....,..,
(!)
o . ....J -2.00
-2.50
- VON KARMAN SPECTUM (13)
L , I
-3.0.?3~OO -2.00 -1.00 0.00 1.00 2.00 3.00 4.00
Figure 43. LOG(nL~/U)
Logarithmic u-Spectra for Run 19, cr and LX Obtained in the High Frequency Range, LX Obtained Via the Direct MMthod u
u
...... ~ ~
0.00 I I
-0.50~ ~ ...
~ -1.00 N~ 0
........ -c I I (:J -cl -1.50 c ~
(!)
9 I I \. :>{+ T ~~~
:.~ .. -t .. ~ .......
-2.00
-2.50
- VON KARMAN SPECTRUM (14)
-3.00 ' " -3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 4.00
Figure 44.
LOG{nL~/U) Logarithmic w-Spectra for Run 19, a and LX Obtained in the Middle-Frequency Range, LX Obtained Via the Direct M~thod w
w
I-' +:-VI
0.00-· --~--------~----------------------I
-0.50 I ~ 1-
x+: I t
~
~ -1.00
-! "'"
II!)A -c: I (!) ;r -1.50 c: ........, (!). / '\.,.~ I-'
.Il-
0 Cf\
..J -2.00
-2.50
- VON KARMAN SPECTRUM (14)
" ~~~~ .. -3.0~3.00 -2.00 - 1.00 0.00 1.00 ·LOG (nL~/U)
2.00 3.00 4.00
Figure 45. Logarithmic w-Spectra for Run 19, cr and LX Obtained in the Righ-Frequency Range, x • w w
L Obtained Via the Duect Method w
E
80· , o.
60 o
I 0
i /
b. / I
/ /
b. /
/ o
o
o ~40 /
/ I 0
I 20 • 0
1° 0
0
A / /
/ /
b. /1 0 /
~/
o
o HOT FILM(VK - HF) A HOT FILM (VK-MF) o CUP_VANE
--ESDU [29] ---~TEUNISSEN [25]
25 50 75 100 125 150 175 200 225 X
LU,m Figure 46. Variation of Average Integral Scales, LX, Obtained with the Hot-Film System
(Von Karman Method, High and Middle Freijuency Range) and with the Cup-Vane System (Data of the Two Systems Taken Simultaneously), with Height for Southerly Winds
I-' ~ '-I
80
60~
E.- 40 N
20
0
0
06
o 6
~
06
l::.
o HOT FILM (VK-HF) l::. HOT FILM (VK-MF)
ESDU [29J I ----TEUNISSEN [25]
// ,,/
...... / -'
"",/
"","'" "","'"
//
//
//
//
0' , o to
Figure 47.
20 30 40 50 60 70 80 90 X
LV,m Variat'ion of Average Integral Scales, LX, obtained with the Hot-Film System (Von Karman Method, High and Middle Frequency Range) with Height for Southerly Winds
..... ~ 00
60 o l::.
E N 40
o l::.
o l::.
20 o HOT FILM (VK- HF)
a HOT FILM (VK-MF)
-- ESDU [29J ---COUNIHAN [26J
- - - - - TEUNISSEN [25J
00 10 20 x
LW,m 30 40
Figure 48. Variation of Average Integral Scales, ~, Obtained with the Hot-Film System (Von Karman Method, High and Middle Frequency Range), with Height for Southerly Winds
I-' ~ \0
80~i----------------------------------------------~ 01
60
E ~40
N
20
0' o
I l:l. 0
I I
/ I I 0 ~ 0
I I
I / I ~
I , I /~ I 0 0 o HOT FILM (VK-MF) I
I I
/ o HOT FILM (DIRECT-MFl I ~ CUP VANE
I
0)' 0 ESDU [29J /
---COUNIHAN [26J D I ----TEUNISSEN [25]
I I
100 200 300 400 x LU,m
Figure 49. Variation of Average Integral Scales, L~, Obtained with the Hot-Film System (Von Karman and Direct Methods, Middle-Frequency Range) and Cup-Vane System with Height for Northwesterly Winds
J
500
t-' \..rI 0
I 8°1 I 0/ . I I
60 ... o
o
I I I
[] I [J
E .. 40 ... I I I N
o / /
[J
/ 201- k /'
0/ / [J
[J
o HOT FILM (VK - MF)
[J HOT FILM (DIRECT-MF) -- ESDU - [29]
I
00
Figure 50.
I
100 ~
200 x LV,m
---- TEUNISSEN [25] I
300
x Variation of Average Integral Scales, t;., Obtained with the Hot-Film System (Von Karman and Direct Methods, Middle Frequency Range) with Height for Northwesterly Winds
400
,~
VI ~
80
GOL
~40 N
20 I " • II
0 r' 0
0
Figure 51.
o HOT FILM (VK -MF) 0
'0 HOT FILM (DIRECT - MF ) ESDU [29]
IW I ---TEUNISSEN [25] 0
o
0
20 40 60 80 100 120 x
Lw,m Variation of Average Integral Scales, LX, obtained with the Hot-Film System (Von Karman and Direct Methods, Middle WFrequency Range) with Height for Northwesterly Winds
...... U1 tv
on / ov I I / 6 o 0 I I
60 o 0
I -I I
I I
6 / I
/
I
E~40 I
/ I
I
/ N
20
00
0/0 /
/ o /
OjID
/
100
~
/
x LU,m
, / , o RUN 16 (VK-MF) o RUN 16 (DIRECT-MF) ~ AVERAGE (VK -MF)
--ESDU [28J ---COUNIHAN [26J --- -- TEUNISSEN [25J
200 300
Figure 52. Comparison of Integral Scales, L~,of Record #16. (Evening Run) (Von Karman and Direct Methods, Middle Frequency Range) Against Average of Daytime Hot-Film Runs (Von Karman, Middle Frequency Range) for North,,,esterly Winds
..... VI W
80 , , I I I
60
E ,.;40'
20
<> O~
I I <> IS)
I I I
I /
<> o~ / / ,I
<> ' / JS)/
~/A:J
OJ j
[] / /
I
[J r o AVERAGE (VK - HF) I
/ o AVERAGE (VK-MF) <> RUN 16 (VK-HF) 6 RUN 16 (VK-MF)' I ,
/ ESDU [29] --- COUNIHAN [26] ---- TEUNISSEN [25] o ' , I '
o
Figure 53.
100, 200 300 x LU ,m
x Comparison of Integral Scales, L\i, Obtained from the Hot-Film System and Analyzed in the Middle-and High-Frequency Range Using the Von Karman Method 'for Northwesterly Winds
~ V1 ~
155
4
~~ , ,'~ ') W~M¥~!/~\llj11\{~i1~#'~¥~~\~~J~¥IN"~~ -2 ~ ~ff~n 'i(1 I lt~' I~ \ I III .
4
2 'cff' ,0 >
4
RECORD OF STREAMWISE VELOCITY COMPONENT, RUN 19,250 ft.-LEVEL, U= 13.9 m/sec , Ciu = 1.60 m/sec,
3-25-17'7, 14:21:00, L~ =306m
II , }" I !1'iv~I~~;I~;t}bi~~I,,~~t~~II., 'IJ;Hl~ 'lifiii ' ,W ' fl'! \ I' !~' I n~"" RECORD OF LATERAL VELOCITY COMPONENT,
RUN 19, 250 ft.-LEVEL, Civ =1.65m/sec, L~:479m
~t~, ~~~IJ~~JI~I~i/lll~ll~fi.!;~I!I!!M!II\·~,'~M~!J~~WI!~d I~ :) ,I. 'I~ . Ifflfrn'l "r/~"r '\1'111\11 1111,\ "IjFI!lp~1 ~, ~'~i! IV : 1 li~'I~,: I
-2 ~ I , \ I! J I/" I i -4 RECORD OF STREAMWISE VELOCITY 'tOMPONENT,
4 RUN 16, 250 ft.- L.EVEL, U = 11.4 m/sec, Ciu = 0.81 m/sec ,
3-23-177, 20:30:00, L~ =77.9m
~~ , ~~~~~~~~~I~~\~r~~)~I~I~~~lif~j~t~II\~~\\~il~~tl#~~ RECORD OF LATERAL VELOCITY COMPONENT,
x RUN 16,250 ft- LEVEL, Ci'v : 0.62 m/sec, Lv=23.3m
Fi.gure 54. Time Records of Velocity Components. Elapsed Time 55 Minutes.
80. I 0 0 I ~ /
/ /
60 J- 00 / A / /
/ /
~ /
~40t- I / /
/ 0 0 / 6. /
,/,
/ o RUN 16 (VK-MF)
201- h / 0 RUN 16 (DIRECT -MF)
/' '0 0 /' 6. AVERAGE (VK - MF)
/ ESDU [29] 0 lI1 / ~
----TEUNISSEN [25] J I
°0 20 40 60 80 100 1·20 X' Lv,m
Figure 55. Comparison of Integral Scales,~, of Record /116 (Evening Run) (Von Karman and Direct Methods,Middle Frequency Range) Against Average of DC3;y Time Hot-Film Runs (Von Karman, Middle Frequency Range) for North,v€sterl,y Winds
..... V1 (j'\
,..." 'I · QUI 06 0 / 0 ;'
GOL OA o o /
/ /
E N40
20'
¢o
~o
~ 0
o o
o
//
// ,..,.
/ /
/
/ /
/
/ /
/
o AVERAGE(VK-HF) o AVERAGE (VK -MF) <> RUN 16 (VK-HF) ~ RUN 16 (VK-MF)
--ESDU [29]
o' " ---TEUNISSEN [25]
o 20 40 x60 80 100 120 LV,m
Figure 56. Comparison of Integral Scales, LX, Obtained from the Hot-Film System and Analyzed in the Middle-and High-F¥equency Range Using the Von Karman Method, for Northwesterly Winds
I-' U'1 -...J
80r'---------------------------------s--o o
60
E .. 40 N
20
0 0
Figure 57.
o
10
IJ
x LW,m
o o RUN 16 (VK-MF)
o RUN 16 (DIRECT-MF) 6 AVERAGE (VK-MF)
--ESDU [29] ----TEUNISSEN [25J
20
'x Comparison of Integral Scales, r,;, of Record 1116 (Evening Run) (Von Karman and Direct Methods, Middle Frequency Range) Against Average of Day Time Hot-Film Runs (Von Karman, Middle Frequency Range)
30
6
I-' \J1 00
60 ~ o
~40 - 0
N
I o ~/~ <:> ~ .------- ---o AVERAGE (VK-HF) o AVERAGE (VK-MF)
20 r //r I I o RUN 16 (VK-HF)
A RUN 16 (VK-MF)
~. ESDU [29] ----TEUNISSEN [25]
Ii
00 10 x20 30 40 Lw,m
Figure 58. Comparison of Integral Scales, ~, Obtained From the Hot-Film System and Analyzed in the Middle-and High-Frequency Range Using the Von Karman Method for Northwesterly Winds
~ V1
'"
160
lOO------~--------------------------~ I' 1 '\ \ I ,
I ' I
Zl
10-1
f=f N 0
b~ "--c: -~ en c:
10-2
- VON KARMAN SPECTRUM (16)
10-3~----~----~------~----~------~
10-1 100 10' 10 2 103 104
fIfo Figure 59a. Logarithmic u-Spectra Versus Modified Reduced
Frequency, flf , Analyzed in the High-Frequency Range, South W~nds, Unstable Thermal Stratification. -1. O<z/L<O
161
100 .___---r-------'--------,
10-1
01
b:3 "'--c .......
:3 en c:
10-2
'1\ I ' I ' , " , ,
I I
f=fo I I , I I
-- VON KARMAN SPECTRUM (16)
IO-3~----~------~------~----~-----~
10-1 100 101 102 103 104
fIfo Figure 59b. Logarithmic u-Spectra Versus Modified Reduced
Frequency, flf , Analyzed in the High-Frequency Range, South W~nds, Stable Thermal Stratiffcation. O<z/L<+1.0
162
100 ------------------------------------
10-1 I I I
f=fo
VON KARMAN SPECTRUM (16)
lo-a ~-----~~-----~----~------~-----~ 10-1 100 101 102 loa 104
fIfo
Figure 59c. Logarithmic u-Spectra Versus Modified Reduced Frequency, flf , Analyzed in the High-Frequency Range, South W~nds, Extremely Stable Thermal Stratification. z/L>+1.0
163
IOO~-----~----------------------------~
01 >
.t~ ....... c: -" >
CI.) c:
I
I', I ' I '\ I ' I -
I I
f=f o . I ..
- VON KARMAN SPECTRUM (16)
IO-~-----~------~------~-----~----~ 10' 10 2 10 4 100
fIfO
Figure 60a. Logarithmic v-Spectra Versus Modified Reduced Frequency, flf , Analyzed in the High-Frequency Range, South W~nds, Unstable Thermal Stratification. -1.0<Z/L<O
164
100~----~------~--------------------1\ 1 ,
1 ' 1 " I I
10-1 I I I I '" t?
"--c: -> en c:
f=f I 0
I I I I 1 I I I I I I I I I I I I I I VON KARMAN SPECTRUM (16) I
10-1 100 10' 10 2 103 10 4
fifo
Figure 60b. Logarithmic v-Spectra Versus Modified Reduced Frequency, f/fQ' Analyzed in the High-Frequency Range, South W1nds, Stable Thermal Stratification, O<z/L<+1. 0
01
~ "'-,..... c ~
> en c
165
100------------~--------.---------------~
10
I' 1 " I \ I \ I ~&A.lI""
I I I
f=f I 0
1 I 1 I I I I I I I I I I I I l - VON KARMAN SPECTRUM (16)
IO-3~----~--.----L------~-------~----·~
10-' 10 0 10' 10 2 10 3 104
fIfo
~igure 60c. Logarithmic v~Spectra Versus Modified Reduced ~requency, flf , Analyzed in the High-Frequency Range, South W~nds, Extremely Stable Thermal Stratification, z/L>+l.O
01
~ "--c: -~ (J) c:
166
100~ __________________________________ ~
10-1
1\ I \ I \ I \ I I 1
I f=f
I 0
I I I I I 1
10-2 ,I I , , 1 I , I I I I
VON KARMAN SPECTRUM (16)
10-~~----~~----~----~------~----~
10-2 10-1 100 10' 10 2 lOa fIfo
Figure 61a. Logarithmic w-Spectra Versus Modified Reduced Frequency, flf , Analyzed in'the High-Frequency Range, South W~nds, Unstable Thermal Stratification, -1.0<z/L<O
.-c
167
100------~----.----------------------~ , ,
10·-'
, , , ,
VON KARMAN SPECTRUM (16) 10-:
3 ~ __ --'-_____ ...I..-___ ..L ___ ....I-.. __ --1
10-1 100 10' 10 2 103 10 4
Figure 6lb. Logarithmic w-Spectra Versus Modified Reduced Frequency, f/f , Analyzed in the High-Frequency Range, South W~nds, Stable Thermal Stratification, O<z/L<+l.O
'"
168
100------~----------------------~
10-1
1\ I ' I " I , I ' I
I I
t:! 'k- f=f I 0
" -c: -~ en c:
I I I I I I I I I I J
I I I I VON KARMAN SPECTRUM ( 16)
10-3~----~----~----~----~----~
10-1 100 101 102
Figure 61c. Logarithmic w-Spectra Versus Modified Reduced Frequency, f/f , Analyzed in the High-Frequency Range, South W~nds, Extremely Stable Thermal Stratification z/L>+l.O
10 4
60
E 40 ... N
o
20
00
Figure 62.
o
0 (fo>U 6 (fo)V 0 (fo)W
I I I I
0.1 0.2 0.3 0.4 fo=( nz/ U)o
The Vertical Distribution of the Reduced Frequency (f ). with i=u,v,w for Southerly Winds. Data Analyzed in 0 1
the High-Frequency Range 0.0244<n<100 Hz.
0.5
...... (j'\ \C
170
,OO~----------------------------~
10-1 ~
o 0
00
o 10-2~
10-3~------~'------_'------__ '----------'-----.J . 10-3 10-2 10-1 100 10' 10 2
f = nz/U Figure 63. Logarithmic u-Spectrum Versus Reduced Frequency,
Analyzed in the Middle-Frequency Range, South Wind,Record #7, z=9.14m
01
b::J "'--c ........
::J (f) C
171
IC)O ...... --------,-------~
o 0
I C)-I
($)0
00 ~O o a (J)
0 0 ,0 0 o
CO
\ 10.-2
10.-3 '----....&..----..101-. ---..... -------'-----'
10-3, 10-2 10-' 100 101 10 2
f = nz/U
Figure 64. Logar:f.thmic u-Spectrum Versus Reduced Frequency, Analyzed in the Middle-Frequency Range. Southwind, Record #7, z=4S.7m
100.
10-1~
C\J
~ "--c: ........
:l (/) c:
-2 10
-3 1
10 10 3
Figure 65.
-I-X I '\
X '\ 6. +CA ~ '\ X
I ~~'\ (!) ~6. '\ ~ eJl~ ~»~.'\
6- bX I 81 ... 1
e:l ~ )
I I I I
f=fo I
~~ I I 1 RUN 19 I 0 15.2m I 6- 30.5m I + 61.0m I x 76.2 m I von KARMAN SPECTRUM I
-2 10 10-1
100 10 I 10
2 10
3 10
4
f/f 0
Logarithmic u-Spectra Versus Modified Reduced Frequency, flf , Analyzed in the High and Middle Frequency Range, Northwest Wind, ~=30lo 0
I-'
" N
-I 10
N> b "-
I--c _. >
(f) c
-2 10
Figure 66.
f -I -2 Ie 10
f=fo I I I I I 0
I f:!,.
I + I x
I I I
fifo
RUN 19 15.2m 30.5m 6LOm 76.2m von KARMAN SPECTRUM
Logarithmic v-Spectra Versus Modified Reduced Frequency, f/f , Analyzed in the High and Middle Frequency Range, Northwest Wind, ~=30lo 0
I-' ...... w
C\J~ b ~ c -~
(f) C
100·--------TK~----------I', !, I p '. .~~"., X
~~'\r
10' + ><,6. ~ Cl~ l:tJ CJ
.6 (!)
I (!)
CJ 1
I f=fO
I ~~
1
102 I 1 RUN 19 I 0 15.2m I A 30.5m I + ·61.0 m I x 76.2m I von KARMAN SPECTRUM
-3· I I 10
10-3 10-2 10-1 10° 101 102 " 103 104
fifo Figure 67. Logarithmic w-Spectra Versus Modified Reduced Frequency, flf , Analyzed in
the High and Middle Frequency Range, Northwest Wind, ~=30lo 0
I-' '-I ~
H
b~
" -c -elf' c
lcol , \ 11\ I
i ~ + I .
I 10-1
I
10-3 .
Figure 68.
6. ...
Cl
fa
~
10-2 10-1
; ~ +-
X fa ~~_ b.~ 'Of~
'" Q1 ~+ X
X ~
I fffo I
RUN 16·
L:i 9.i m A 15.2m. + 30.5m X 61.0 m ~ 76.2 m
---VON KARMAN SPECTRUM
'100
fIfo 101 102 '03 104
Logarithmic u-Spectra Versus Modified Reduced Frequency, flf , Analyzed in the High and Middle Frequency Range, Evening Run, Northwest aind, ~=292°
.... ~ VI
10
10-1. tid - I '"
t? ~ ~~ ~ f=f ....... f:l <Ix I 0 -c
--- 6. > "t-
en ~ CJ~ C .... , '-tV' X t-' -..J 0'1
10-2r r ' .. i~."'-. x
RUN 16
[!) 9.1 m A 15.2m
+ 30.5m
I x SO.lm ¢ 76.2m ,-VON KARMAN SPECTRUM
I ,
10-10
-3 10-2 10-' 100 10' 102 103 104
fIfo
Figure 69. Logarithmic v-Spectra Versus Modified Reduced Frequency, flf , Analyzed in the High and Middle Frequency Range, Evening Run, Northwegt Wind ¢=292°
-~K\-'------ I 10°1-- 1\
I 1~~ 10-1
N ~ 6 a 9 I ~ "'- ~ f=f -c: 8 . 0 - ~ I
~.
CJ~ / CJ)
'~. V
I-' c: ......
I ......
10-2r I. RUN 16 I""l I ~. -
Cl VI II!] 9.1 m
A 15.2m
I + 30.5m x 61.0 m ~ 76.2m
1- VON KARMAN SPECTRUM
10-IO-~ 10-2 10-1 100 101 102 loa 104
fifo Figure 70. Logarithmic w-Spectra Versus Modified Reduced Frequency, flf , Analyzed
in the High and Middle Frequency Range, Evening Run, Northwegt Wind, ¢=292
i 80, o~ o • • •
C601-0 ~. 0 & •
E 401-... N
~O~
'20 rO~
00
o
'0
&
.&
.& 0
• o • (fo)u
• It:.. Ho)v • 10 • (fo)w
.0"" -----~----- I I o 0.1 0.2
fo=(nz/U)o 0.3
Figure 71. The vertical Distribution of the Reduced Frequency (fo) i Withi=u,v,w for Westerly Winds. Data Analyzed in the Middle-Frequency Range O.0015<n<6.25 Hz. Dark Symbols Denote the Data. for the Evening Run 16.
.....
....... 00
o It)
1.0 I I I I . I I I I I
. ~J----r1 ~------------ -_I l· l -O.~O. ... 1 @-
-J----~-1------l----------I- ~--- --- . r-------1----- I- e· __ •
~--------------
0~90 ~ ~-. ~--.
I--1----
0.85 ~ I I I
.. ... .. .. .. .. >(\J 0.80 v .. ~
l- f-*- .>L
~ =15.2 m r--3> )( )(
~.= 30.5m
0.75 .. ----- ! = 45.7 m -.-._- ~ = 61.0 m
0.70 .,
~
.. ~
+ .. 0.65
0 • • • . ..
60 120 180 240 300 360 WIND DIRECTION, DEGREES
Figure 72. .Average Velocity Ratios, V/V250 for Each Wind-Direction Sector with NearUniform Upstream Roughness (See Figure 3)
...... ....... \0
24
20
16
o ~ 12 ::J "-b::J
8
4
~ .. I-
Mt-
t--
t--
---_. 1-'-_.-._. r-"--~.- ..
-
I-
00
Figure 73.
)( )( z=15.2m
z= 30.5 m
---- ~ = 45.7 m """""'" v .. n .. .. . .. ~
------ ~ = 61.0 m n n n
_ .. _-.- ! =76.2m
l-
t---- f::".: :--~:. =.-:. 1----
HI- F:::.::-:-.:. --- 1-"-"-"-" r-'-'-' _._. 1-"-"-_ .. -. .. .. .. .. ..
---__ " "'- _._.
r-"-.
1------- - - -----~.-.-.-.-.-.-.-.-. 1- •• - •• _ •• _ •• _ •• _ •• - •• - -----
1-'-'. 1"'"-' t- .. - ..
~:::
I J I
60 120 180 240 300 360 WIND DIRECTION, DEGREES
Average Turbulence Intensity, cr Iu, for Each Wind-Direction Sector with Uniform Upstream Roughness (SeeUFigure 3)
I-' 00 o
~ o .. :::) ........
t?
18 )( X i! = 15.2 m .. ...
. - i! =·30.5m .. ~ ~
n
i! = 45.7 m .. ---- ,-~ _a ___
~ = 61.0 m
i!- = 76.2 m ~----
f-+ --._ ... ------- _._. _. _ .. _ .. -
1-----1--- ~:-:--=--;0: :.= ••• -== I-- t-*- 1= :::.:.::-:. 1-._. ... ---- 1- .. -
~.-.-.. I .. .. I I I I I I I ~ .. - .. - 1---
12
1--- .. .. ~'-' . F-'.-:-~~ .. -.
6 ~ 1----- - --- - -- ----~.-.-.-.-.-.---.-j.-. •• _ .•• - •• _ •• - •• _ •• - ••. ~.
t---1-----1---1--'-'-1--"-" F=:.":
• I I • 00 60 120 180 240 300 360
WIND DIRECTION, DEGREES Figure 74. Average Turbulence Intensity, 0 /U, for Each Wind-Direction Sector
with Near-Uniform Upstream RougKness (See Figure 3)
I-' 00 I-'
350 )( )( ~ = 15.2m
1-"-" ~
i! = 30.5 m
---- i! = 45.7 m 1-'- .-
300 _.-.- i! = 61.0 m 1-',-"-,,-,, !-
1-._._._._. - .. _ ... ~ = 76.2 m 1-------
250 I=:~::.~ 1- •• _. ,..--_. 1-----t-- •• - •• - • -""- •.• - •• - •• _. 1-'-'
t---~.-.-.-.- ._.-._._.- I- •. - •• -
E.- 200
,.. .. - ,.. ... ~.-.-.~ ..... - r------------ -----
x=, _ .. _ ...
-.J -._.--- .-
r"-' ~V .. ." ---- ~.-.
150 -'!L -'!L ~ t--- ----
r--- --- v v
100 ~ v .. .. .. --"'-
r-" OIL _v
v
~ 50
0 I I .1- I
60 120 180 240 300 360 ~ WIND DIRECTION, DEGREES
Figure 75. x Average Turbulence Integral Scales, L , for Each Wind-Direction Sector with Near-Uniform Upstream Roughness u(See Figure 3)
i--' co N
TABLES
TABLE I. Mean Profile and Turbulence Parameters for Strong Winds Over Near-Uniform Terrain
Cup-Vane Instruments
li V~2 IU* LX(m) * Data <P (m/s) U z Ri Type of Date Time EST Record (deg) 0 u 0 0 a Profile or EDT at z= at z= at (m/s) (m) at z=
15.2m 15.2m 76.2m 15.2m
158C 241.4 13.5 2.83 303 0.81 0.019 0.131 -0.026 S 2-25-'77 12:37-13:45 158D 244.0 12.9 3.10 149 0.78 0.020 0.137 -0.026 S 2-25-'77 14:02-15:02 158E 247.0 11.6 2.84 201 0.77 0.037 o .1lI6 -0.023 S 2-25-'77 15:02-15:53 153B 288.3 13.3 2.98 393 0.86 0.031 0.164 -0.027 S 12-21-'76 12:22-13:39 153C 291.1 12.5 3.15 420 0.81 0.031 0.146 -0.028 S 12-21-'76 14:30-15:38 153A 291.9 13.2 3.08 429 0.85 0.031 0.140 -0.028 S 12-21-'76 11 :14-12:13 168D 297.9 11.4 2.50 342 0.80 0.052 0.157 -0.053 K 3-25-'77 14:12-15:46 168C 312.2 11.2 2.69 639 0.73 0.034 0.150 -0.060 K 3-25-' 77 12:38-13:38 128G 316.2 16.8 2.80 191 1.04 0.023 0.136 -0.016 S 8-9-'76 17:07-18:07 f-"
00· 128H 316.6 13.9 2.80 125 0.92 0.037 0.150 -0.021 S 8-9-'76 18:07-18:50 +:-
l}
c 144A 317.8 10.9 3.15 167 0.65 0.018 0.140 -0.044 S 10-14-'76 9:58-10:58 en 169C 319.2 13.1 2.39 319 0.98 0.073 0.173 -0.029 K 3-31-' 77 13:27-14:27 G) 0 163A 323.0 10.6 2.64 201 0.69 0.030 0.152 -0.046 K 3-16-' 77 10:53-12:01 < m 169D 323.0 12.2 2.70 236 0.87 0.055 0.163 -0.035 K 3-31-'77 15:09-16:09 jJ z 128F 340.0 17.1 2.50 209 1.16 0.040 0.160 -0.082 K 8-9-'76 15:08-16:25 s: m Z -I
" Hot-Film Anemometers jJ
Z 15 294 13.0 2.38 433 0.92 0.082 0.20 -0.039 K 3-23-'77 15:20-16:15 -I Z 14 296 12.0 2.79 426 0.85 0.070 0.17 -0.040 S 3-23-'77 13:30-14:25 G)
0 19 301 11.6 2.57 306 0.78 0.040 0.16 -0.046 K 3-25-' 77 14:21-15:16 "T1 "T1 17 304 12.5 2.61 541 0.87 0.049 0.15 -0.041 S 3-24-' 77 12:27-13:22 () ~ 18 322 11.9 2.36 397 0.90 0.073 0.18 -0.025 K 3-24-' 77 16:25-17:20 U; 16 292 7.1 1.03 78 0.94 0.741 0.25 -0.0014 S 3-23-'77 20:30-21:25 00
.!oJ ~
~ Note: Type of Profile: S-sing1e logarithmic profile, K-kink in profile. ~ en "-~ co N
End of Document